context stringlengths 100 2.82k | A stringlengths 100 3.62k | B stringlengths 100 3.46k | C stringlengths 100 5.23k | D stringlengths 100 6.82k | label stringclasses 4
values |
|---|---|---|---|---|---|
For this data, we downloaded the bpRNA dataset from SPOT-RNA at https://sparks-lab.org/server/spot-rna/, bprna-1m data from https://bprna.cgrb.oregonstate.edu/ and used RNAStralign, based on E2Efold, from https://github.com/ml4bio/e2efold. | For RNA MSA construction, we built the database using a nucleotide database (ftp://ftp.ncbi.nlm.nih.gov/blast/db/FASTA/nt.gz), Rfam (https://rfam.xfam.org) and RNAcentral (https://rnacentral.org) and use rMSA (https://github.com/pylelab/rMSA) for searching and construction tools. | j. Violin plot of RhoFold+’s RMSD values in the cross-family validation. Here, all the structures in a family to be tested were masked during model training, and RhoFold+ accurately predicted RNA structures from most unseen families. The numbers of sequences in each family are shown in parentheses. | i. Overview of cross-type validation performance of RhoFold+ measured by lDDT and TM-score. All structures in the type used for validation were masked during model training. | The family/type information in Rfam (https://rfam.xfam.org) was used for cross-family/type validation. | D |
Equation (12) shows that, for a certain form of maternal effect, there is a natural relationship between the time derivative of (the logarithm of) the trait value and the logarithm of the population size. | The momentum-averaged positions 𝔼(𝐳(t))𝔼𝐳𝑡\mathbb{E}({\bf z}(t))blackboard_E ( bold_z ( italic_t ) ) and velocities ddt𝔼(𝐳(t))𝑑𝑑𝑡𝔼𝐳𝑡\frac{d}{dt}\mathbb{E}({\bf z}(t))divide start_ARG italic_d end_ARG start_ARG italic_d italic_t end_ARG blackboard_E ( bold_z ( italic_t ) ) are discrete analogs of mome... | By choosing the form of the selection “force” via f(z(t))𝑓𝑧𝑡f(z(t))italic_f ( italic_z ( italic_t ) ), we can consider different kinds of “bound motion” of which simple harmonic motion is a fundamental example. | The Price equation [36] pertains to selection in evolutionary processes. It was motivated by a desire to understand the evolution of altruism [22], and has been described as a “fundamental theorem of evolution” [37] due to its generalization and unification of many results in evolutionary biology. For example, Fisher’s... | Note that this is the second order differential equation describing simple harmonic motion (SHM) for the logarithm of the trait, where k𝑘kitalic_k is the analog of a spring constant. The “stiffness” of the spring, k𝑘kitalic_k, is related to the strength of the trait’s effect on fitness. The angular frequency of the m... | B |
This indicates that using these regressors for overestimation prediction may not always enhance solution sets, | Interestingly, dummy and SVR-based MO optimizers consistently outperform the CHV of unadjusted models, | for the regularization strength of the SVR with random search, utilized for predicting overestimation, underperformed with respect to dummy and SVR regression models. | weighted median (dummy), pruned decision tree (ptree), random forest (RFReg), support vector regression (SVR), and SVR with optimized regularization parameters (rSVR). | Our study found that even a basic regression model, which learns solely the weighted average of overestimation and is used for fitness adjustment, resulted in improved overall performance compared to the unadjusted optimizer in 7 out of 8 experimental setups, as indicated by the CHV metric. Likewise, the MO optimizer d... | A |
In the past few years, state-of-the-art, DNN-based methods were introduced for IVIM parameter estimation. Bertleff et al. [8] demonstrated the ability of supervised DNN to predict the IVIM model parameters from low SNR DWI data. Barbieri et al. [6] proposed an unsupervised, physics-informed DNN (IVIM-NET) with results ... | We address the challenge of estimating the IVIM model parameters while compensating for motion artifacts by presenting a self-supervised DNN-based framework for simultaneous motion compensation and IVIM model parameters estimation. | We evaluated the anatomical registration accuracy of our IVIM-Morph in comparison to the different registration approaches outlined in Section 4.2 for cases with different levels of motion. These techniques were utilized to assess the alignment of images Sisubscript𝑆𝑖S_{i}italic_S start_POSTSUBSCRIPT italic_i end_POS... | Nevertheless, all these algorithms presuppose spatial alignment among the different b-value images, rendering them unsuitable for direct application in estimating IVIM model parameters for fetal DWI data, given the inevitable fetal motion during acquisition [1]. | In the context of abdominal imaging, [29] introduced an iterative motion correction model to address the differences in image contrast in the DWI images by registering images and estimating parameters with the IVIM model. Similarly, [35] simultaneously compensates for motion and performs qDWI analysis using a mono-expo... | C |
At a time t=15001𝑡15001t=15001italic_t = 15001, while the agent population has mostly completed learning and some agents have accumulated at the target (Fig. 2 (a) and (b)), the single agent is still exploring the target. | Furthermore, the value distribution becomes irregular (Fig. 4(a) t=1001𝑡1001t=1001italic_t = 1001) due to random exploration histories and the lack of a diffusion term. | In this one-trial comparison, the overall update speed of Ztsubscript𝑍𝑡Z_{t}italic_Z start_POSTSUBSCRIPT italic_t end_POSTSUBSCRIPT by the single agent is slower than the agent population due to its limited exploration range (Fig. 4(a) t=1001𝑡1001t=1001italic_t = 1001). | We observe that the agents expand their exploration area by dispersing at branching vertices and shaping a gradient toward unexplored directions via degradation of the endogenous cue (Fig. 2 (a) and (b) at t=21𝑡21t=21italic_t = 21 and t=1001𝑡1001t=1001italic_t = 1001). | During transient states, the agents produce the endogenous cue faster than agent following the optimal log-exp coupling, despite both following an exponential production dynamics (Fig. S1 (d) t=1001𝑡1001t=1001italic_t = 1001). | A |
In the state predictive information bottleneck (SPIB) extension of IB [45], the inputs are molecular features at time t𝑡titalic_t, 𝐡~(𝐗t)~𝐡subscript𝐗𝑡\tilde{\mathbf{h}}(\mathbf{X}_{t})over~ start_ARG bold_h end_ARG ( bold_X start_POSTSUBSCRIPT italic_t end_POSTSUBSCRIPT ), and the targets are state labels stsubs... | The latent representation and state labels are learned simultaneously by predicting the state labels at time t+τ𝑡𝜏t+\tauitalic_t + italic_τ given the molecular features at time t𝑡titalic_t. | In the state predictive information bottleneck (SPIB) extension of IB [45], the inputs are molecular features at time t𝑡titalic_t, 𝐡~(𝐗t)~𝐡subscript𝐗𝑡\tilde{\mathbf{h}}(\mathbf{X}_{t})over~ start_ARG bold_h end_ARG ( bold_X start_POSTSUBSCRIPT italic_t end_POSTSUBSCRIPT ), and the targets are state labels stsubs... | In this work, we prepared the initial state labels by performing k-means clustering on the CVs learned from VAMPnets based on distances between Cα atoms with k=100𝑘100k=100italic_k = 100 clusters. Training reduces the number of distinct states that are populated to the estimated number of metastable states. Further tr... | where χ0subscript𝜒0\chi_{0}italic_χ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT and χτsubscript𝜒𝜏\chi_{\tau}italic_χ start_POSTSUBSCRIPT italic_τ end_POSTSUBSCRIPT are vectors of functions and the expectation is over trajectories initialized from an arbitrary distribution μ𝜇\muitalic_μ. In our case, 𝐗tsubscript𝐗𝑡\ma... | A |
Pranav Verma: Coding, Methodology, Formal analysis, Writing – original draft. Viney Kumar: Coding, Methodology, Formal analysis, Writing – original draft. Samit Bhattacharyya: Conceptualization, Methodology, Formal analysis, Writing – review & editing, Supervision. | Our research suggests that the public health intervention of providing hospitals for quarantine has a favourable effect on the quarantine and hospital burden if the disease transmission rate is high in the community. Additionally, this can help reduce the prevalence of disease in the population. During a pandemic, publ... | We also analyzed the difference between the baseline payoff and the new payoff of the disclosure without 1−ηq1subscript𝜂𝑞1-\eta_{q}1 - italic_η start_POSTSUBSCRIPT italic_q end_POSTSUBSCRIPT to determine the importance of public health measures (Figure 9 &\&& S2). When ηqsubscript𝜂𝑞\eta_{q}italic_η start_POSTSUBSCR... | The authors thank all reviewers for their insightful comments and suggestions for improving the presentation of the content in the manuscript. This work was initiated under the Shiv Nadar Institution of Eminence (SNIoE)’s Opportunity for Undergraduate Research (OUR) program. Viney Kumar thanks the Council of Scientific... | However, no study considers disclosing exposure to infection as an individual choice in the face of an outbreak and limited availability of medical facilities. Understanding the interplay between individual strategies, medical facilities’ availability, and its consequences on disease burden is an important challenge fo... | C |
=β~exp(S∗+D∗−I∗)−γ,absent~𝛽subscript𝑆∗subscript𝐷∗subscript𝐼∗𝛾\displaystyle=\tilde{\beta}\exp(S_{\ast}+D_{\ast}-I_{\ast})-\gamma,= over~ start_ARG italic_β end_ARG roman_exp ( italic_S start_POSTSUBSCRIPT ∗ end_POSTSUBSCRIPT + italic_D start_POSTSUBSCRIPT ∗ end_POSTSUBSCRIPT - italic_I start_POSTSUBSCRIPT ∗ end_P... | Although (II-B) is derived from the mean-field approximation of a Markov process evolving on an infinite random Poisson degree graph, it also serves as a basis for approximate inference regarding both the parameters of the underlying network and the SIR process. This method, which employs ODEs to characterize a stochas... | Note that the equations above may be interpreted as the mean field approximation to the scaled SIR Markov process evolving on the configuration model random graph 𝒢(n,p)𝒢𝑛𝑝{\cal G}(n,p)caligraphic_G ( italic_n , italic_p ) as described in the beginning of this Section. | The rest of the paper is organized as follows: In Section II, we provide background on network-based stochastic SIR models, focusing on configuration models, their pairwise representations, and dynamical survival approaches to model fitting. Following [7], we argue that in some cases the classical SIR model serves as a... | The probabilistic DSA model used to derive (11) and (12) can now be applied in a similar manner to derive the likelihood function for inference of the parameters vector θ𝜃\thetaitalic_θ given in (13). The justification for this approach is discussed in [17] through the so-called Sellke construction. We omit the detail... | A |
Experimentalists have identified interneurons that are primarily responsible for driving and facilitating the primary locomotion behaviors of forward crawling, reversals, and turns [50, 45, 33]. | Individual neurons that promote forward and reversal movements have been identified, but the mechanisms that govern their collective activity remain to be understood. | Assuming that a high state of the forward cluster represents forward locomotion and the analogous statement for the reversal cluster, Figure 3(b) shows the forward/reversal direction of movement for this C. elegans as a function of time. | Ablation studies show that key premotor neurons are essential for producing forward and reversal movements [4, 48]. | For example, we predict inhibitory synapses from forward neurons AVB and RIB to reversal neurons AVE and AIB, and mostly excitatory synapses from the reversal neurons to other reversal neurons, | C |
We have performed overground gait transition experiments over much longer distances than earlier and shown that humans use remarkably similar behavior to short distance bouts. Specifically, we find a gradually morphing walk-run transition regime, that involves walk-run mixtures, dominated by walking at lower speeds and... | Aside from instantaneous energetics, gait transition from walking to running has been attributed to muscle force-velocity behavior [31], interlimb coordination variability [32], mechanical load or stress [33, 18], and cognitive or perceptual factors [34, 35]; see [36] for a review. However, none of these factors can sh... | Most humans do not spend their lives on treadmills, so their behavior may not already be energy optimal for such gait transition tasks without considerable learning [19, 20, 21]. Here, in contrast to these treadmill gait transition experiments, we show that overground gait transitions in realistic overground locomotion... | We have performed overground gait transition experiments over much longer distances than earlier and shown that humans use remarkably similar behavior to short distance bouts. Specifically, we find a gradually morphing walk-run transition regime, that involves walk-run mixtures, dominated by walking at lower speeds and... | Subjects used a mixture of walking and running in 90% of the trials at the two intermediate speeds (2.22 and 2.6 ms-1). On average, walking dominates the walk-run mixture at the lower speeds and running dominates the walk-run mixture at the higher speeds (Figure 2a), so that the walk-run mixture gradually changes as sp... | A |
Van der Zee and Kuo [28] proposed a model of metabolic rate proportional to the second derivative of force, which is equivalent to the metabolic cost per movement being proportional to the first derivative of force. This is a different cost from our model, which they supported by showing an approximate quadratic scalin... | We found that decreasing the force is more costly than increasing the force by having different coefficients in the model for positive and negative force rate (3). One reason positive and negative force rate may have different costs may be due to decrease force, the calcium needs to be pumped back to the sarcoplasmic r... | Here, we focus on developing a metabolic cost model applicable to isometric tasks involving arbitrary time-varying force production based on joint torque and torque rate, which includes constant force as a special case. In previous work, we showed that the metabolic cost of near-constant isometric force scales non-line... | In the main manuscript, we expressed the metabolic cost as a function of external force and force rate, using a single-link model. Now, we consider a limb with multiple joints and multiple muscles. As in our experiment, this limb at rest needs to produce a one-parameter family of external forces and force rates, all al... | Figure 3: Metabolic model. A) Best-fit force-dependent of metabolic cost is slightly nonlinear (exponent γ1=1.4subscript𝛾11.4\gamma_{1}=1.4italic_γ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = 1.4). The curve shows model metabolic cost when the force is constant and force rate is zero. B) Best-fit force-rate-dependent of... | A |
Here we assume that each secondary infection differs slightly from its infection source. For the analytical calculations as well as the numeric, we employ a standard SIR model with minor adjustment - the recovery population is also divided to recovered nodes who became healthy and immunized and recovered nodes who died... | We define three main parameters to describe the pathogen – the average time to infect a neighbor node (λ𝜆\lambdaitalic_λ), the average time to recover and get healthy (r𝑟ritalic_r), and the average time for an infected node to die from the disease (or to get recovered in the SIR model also) (γ𝛾\gammaitalic_γ). The γ... | Another advantage of the square grid graph is the ability to easily visualize the graph’s structure and better understand the flow of the epidemic in it. Figure 8 displays the mutation of the pathogen during the epidemic spreading in the network. The α𝛼\alphaitalic_α value of each node is presented as a color in the f... | For the numerical part of this section, we performed numerical simulations similar to those in the previous section. However, in order to simulate the process for random networks, we first generated random networks based on Erdős–Rényi model and simulated the spreading process on them. The simulation starts with a rand... | The mutation process is defined by randomly changing γ𝛾\gammaitalic_γ for each newly infected node, where the change can be in both sides: the new infected node gets a new pathogen with higher or lower mortality mean time. Therefore, the mutated parameter (γ𝛾\gammaitalic_γ, or the mortality mean time) on new infected... | A |
(N∗superscript𝑁N^{*}italic_N start_POSTSUPERSCRIPT ∗ end_POSTSUPERSCRIPT being the formulation used in [19]). | state, r=αN∗𝑟𝛼superscript𝑁r=\alpha N^{*}italic_r = italic_α italic_N start_POSTSUPERSCRIPT ∗ end_POSTSUPERSCRIPT, where N∗superscript𝑁N^{*}italic_N start_POSTSUPERSCRIPT ∗ end_POSTSUPERSCRIPT is the steady-state density | However, the values of αisubscript𝛼𝑖\alpha_{i}italic_α start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT and Ni∗superscriptsubscript𝑁𝑖N_{i}^{*}italic_N start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ∗ end_POSTSUPERSCRIPT need not be | _{n}N_{n}^{*}italic_α start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT italic_N start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ∗ end_POSTSUPERSCRIPT = italic_α start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT italic_N start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ∗ end_POSTSUPERSCRIPT = ⋯ italic_α start_PO... | Since r is a measure of fitness, so must αN∗𝛼superscript𝑁\alpha N^{*}italic_α italic_N start_POSTSUPERSCRIPT ∗ end_POSTSUPERSCRIPT be. | D |
Fig. 6B shows the population average of the steady-state fitness as a function of the number of targeted epitopes, g𝑔gitalic_g, for different values of the effective speed of antigenic evolution, κ=ρτ𝜅𝜌𝜏\kappa=\rho\tauitalic_κ = italic_ρ italic_τ. We find a link between immune recognition complexity and target mut... | This recognition function is well supported by empirical findings. First, clinical studies show that human immune protection correlates well with overall antibody titers in serum, independently of the epitope distribution of the response [79, 80]. Second, functional binding of a single epitope (monoclonal response) is ... | The following example shows how selection on complexity can act in the adaptive immune system, a rapidly evolving recognition system of high global complexity [62]. In the presence of an antigen, immune B cells produce neutralizing antibodies that bind to specific target sites on its surface (called antigenic epitope s... | Fig. 6B shows the population average of the steady-state fitness as a function of the number of targeted epitopes, g𝑔gitalic_g, for different values of the effective speed of antigenic evolution, κ=ρτ𝜅𝜌𝜏\kappa=\rho\tauitalic_κ = italic_ρ italic_τ. We find a link between immune recognition complexity and target mut... | In summary, we have shown that fast antigenic drift can induce selection for complex adaptive immune responses, in line with our general picture of complexity in dynamical recognition systems. This finding is consistent with observations: several fast-evolving RNA viruses, including influenza [25, 82], norovirus [69], ... | D |
Next, we considered the relationship between TLL1 and the maturation process of TGF-β𝛽\betaitalic_β. TLL1 has two mRNA isoforms: TLL1 isoform 2 lacks many exons from the 3’ end of TLL1 isoform 1. | To investigate this possibility, as shown in Fig. 11a, we performed a luciferase assay using the pre-mature form of TGF-β𝛽{\beta}italic_β co-expressed with either TLL1 isoform 1 or 2 in a HepG2 cell line. | We next considered the gene expression of TLL1 in ATL cells and TLL1’s effect on the maturation process of TGF-β𝛽\betaitalic_β in a HepG2 line, which is a TGF-β𝛽\betaitalic_β responsive cell line(Westerhausen1991, ). | It should be noted that the difference of luciferase activity between isoform 1 (Case 3) and isoform 2 (Case 4) with pre-mature TGF-β𝛽\betaitalic_β approximated the difference between the TLL1-less condition with pre-mature TGF-β𝛽\betaitalic_β (Case 2) and without pre-mature TGF-β𝛽\betaitalic_β (Case 1). Thus, the r... | As shown in Fig. 11b, we found that compared to the sample without TLL1 (Case 2), TLL1 isoform 1 (Case 3) activates the pre-mature form of TGF-β𝛽\betaitalic_β for maturation, but TLL1 isoform 2 (Case 4) represses the maturation. | A |
Experimental studies of the neural mechanisms of coordination between individuals is an emerging field both at behavioral level Guevara et al. (2017); Haken, Kelso, and Bunz (1985) and in the brain activity level, particularly in the context of coordination of movement Moreau et al. (2023) and other aspects of social c... | Brain activity measured at a macroscopic (cm) scale in humans using electroencephalography (EEG) reflects transient quasi-stable patterns that evolve over timeNunez and Srinivasan (2006). An extensive literature characterizes these patterns as functional networks, using correlation, coherence, or mutual information to ... | Coordination and interaction between humans are an essential cognitive function. There has been growing interest in understanding the brain mechanisms that support such social cognition by simultaneous recording of brain activity during interactions between multiple individuals, an experimental procedure known as hyper... | Using a data-driven approach, we define brain states based on the correlations between different brain areas for each subject, allowing us to create symbolic representations of joint brain states. This paradigm enables the exploration of multi-brain dynamics without relying on millisecond time scale connectivity betwee... | In this study we characterized brain states by the pattern of functional connectivity observed in short epochs of EEG signals (2s long with 50% overlap). We obtained an analytic signal using a Hilbert Transform, and for each epoch computed the complex-valued correlation between each pair of analytic signals to obtain a... | A |
Several computational methods have been proposed for protein-RNA binding affinity prediction, including sequence-based and structure-based methods. The sequence-based approaches process the protein and RNA sequence separately with different sequence encoders (Yang and Deng 2019a; Pandey et al. 2024), and subsequently m... | Many efforts have emerged to develop foundation language models to leverage the massive biological sequence data. One of the first papers is ESM-1b (Rives et al. 2021) trained on 250 million protein sequences with a BERT-style strategy. Several other PLMs are proposed and perform well on various downstream tasks(Rao et... | Recently, many protein language models (PLMs) (Lin et al. 2022; Rao et al. 2021) and RNA language models (RLMs) (Penić et al. 2024; Chen et al. 2022) have been developed, most of which utilize a mask language modeling strategy (Devlin 2018) to pre-train the models with massive unlabeled sequences. They’ve shown great p... | Although the current works show the prosperous potential of structure-informed biological language models for interaction tasks, there are still few works combining pre-trained models from different biological domains. Integrating pre-trained models for multiperspective information extraction has received much attentio... | Learning from multiple modals can provide the model with multi-source information of the given context (Huang et al. 2021). Multi-modal learning achieves impressive performance improvement compared to its single-modal counterparts and brings new applications(Luo et al. 2024; Li et al. 2023). Contrastive learning is one... | B |
_{gi})^{2}}RL2 = ∥ bold_z start_POSTSUBSCRIPT italic_r end_POSTSUBSCRIPT - bold_z start_POSTSUBSCRIPT italic_g end_POSTSUBSCRIPT ∥ start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = square-root start_ARG ∑ start_POSTSUBSCRIPT italic_i = 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT ( italic_z start_POST... | We train the combined network (Eq. 1 and Eq. 2) on real images ℛℛ\mathcal{R}caligraphic_R to maximize the log-likelihood of the real data. This training process is performed only once, allowing the network to map image features into a Gaussian latent space that assigns high likelihoods to real image features. Notably, ... | Figure 1: The process of computing RL2. In the training phase, the ResNet-normalizing flow network is trained on the given real (high-quality) images. In the evaluation phase, real and generated (or evaluation) images are passed through the network, and the L2 distance between mean of latent vectors of real and generat... | The computation of L2 distance between two vectors is computationally lighter and faster than the computing Fréchet distance or maximum mean discrepancy between them. Hence, RL2 is lighter, faster and efficient to compute than previous metrics. | We developed a new evaluation metric based on normalizing flow and measures the L2 distances between real and generated features. Our metric has been shown to be monotonic with respect to various noise types found in histopathology, including noise, occlusion, and blur. Additionally, because our metric is lighter, fast... | C |
We also evaluate the probability distributions of ε𝜀\varepsilonitalic_ε for each disease. Figure 5B shows these distributions, indicating that decreasing returns to scale is the most common regime across the seven disease types. The percentage of cities exhibiting sublinear scaling ranges from 66% for pertussis to 95%... | Here we bridge this gap by investigating the effect of inter-city interactions on the association between population size and the number of cases for seven infectious diseases across Brazilian cities. To do so, we use the commuting network among cities as a proxy for inter-city interactions, combined with a general sca... | As previously mentioned, decreasing returns to scale is the predominant response of cities to a proportional increase in both population and commuters. This regime is more common among cities of intermediate size and connectivity within the commuting network. As these cities grow and enhance their connectivity, they ma... | We find that the majority of cities exhibit sublinear scaling, wherein proportional increases in both population and commuters are associated with less-than-proportional increases in disease cases. However, a significant subset of cities (ranging from 5% to almost a quarter of Brazilian cities) exhibits superlinear sca... | Somewhat counterintuitively, negative scaling regimes have been observed in density scaling laws, particularly in the context of housing prices [48, 49], where the price of detached housing decreases with increasing population density at high densities in England. In our study, cities exhibiting negative elasticities a... | D |
To enhance the performance of downstream classification, we created an extended feature set. Binary-encoded pathway interaction data for each gene was sourced from the Comparative Toxicogenomics Database (CTD) [10]. Similarly, chemical interaction data from the CTD was used to create binary features representing the to... | Figure 1: A) Illustration of the GAN-TAT architecture. The upstream module embeds the PIN and generates an extended feature set. The downstream module partitions the dataset and trains classifiers. B) Designs of ImGAGN-GraphSAGE, with graph generator, encoder, and discriminator. | The ImGAGN model integrates a generative adversarial network (GAN) framework consisting of three principal components: a graph generator, an encoder, and a discriminator [13, 28]. This architecture effectively manages imbalanced and sparse networks, making it ideal for this study. The generator creates synthetic nodes ... | GAN-TAT comprises two primary components: an upstream embedding module and a downstream classification module. The upstream module utilizes the ImGAGN-GraphSAGE model for network embedding, which is a supervised learning model independently trained and optimized using the PIN and a label set. After training, the model ... | In the downstream module, a subsampling strategy is implemented to mitigate class imbalance. The complete dataset is partitioned into M folds. For each fold, 80%percent8080\%80 % of the minority (positive) class data points are randomly selected. Additionally, a subset of the majority (negative) class data points is ra... | C |
One of the key advantages of deep learning over traditional modelling approaches is that deep learning does not rely on manually determined features. Instead, the model is able to learn a large set of relevant features automatically during the training process. These features have been demonstrated to perform better th... | Convolutional Neural Networks (CNNs) - CNNs are a class of deep learning models that use the convolutional neural networks as core components of its model. The convolution kernels serve as filters with weights that resemble manually defined kernel filters after training. The use of convolutional filters also reduce th... | One of the key advantages of deep learning over traditional modelling approaches is that deep learning does not rely on manually determined features. Instead, the model is able to learn a large set of relevant features automatically during the training process. These features have been demonstrated to perform better th... | MASIF [31] is one of the first works that attempts to perform convolution on surfaces of proteins for deep learning. In Euclidean space, the shortest distance between two points is a line, whereas for manifolds it is a geodesic, a curvature along the surface. The convolutional filters used in CNNs are replaced by geode... | An example of this is the set of manually defined filters in SIFT [67] for CV applications. Experiments have shown that deep learning architectures are able to learn similar filters in the different layers of their convolution kernels, but in much larger numbers [106]. Because these are learnt automatically during the ... | D |
4RTNI. This dataset was collected as part of the 4-Repeat Tauopathy Neuroimaging Initiative (4RTNI) and used similar MRI acquisition and clinical assessments as the NIFD dataset. This dataset constituted of 59 individuals diagnosed with progressive supranuclear palsy (PSP; age = 70.79±7.65plus-or-minus70.797.6570.79\pm... | PPMI. This dataset constituted of 454454454454 individuals diagnosed with Parkinson’s disease (PD; age = 63.17±9.37plus-or-minus63.179.3763.17\pm 9.3763.17 ± 9.37 years, 174174174174 females) and 133133133133 healthy individuals (HC; age = 61.44±11.45plus-or-minus61.4411.4561.44\pm 11.4561.44 ± 11.45 years, 45454545 fe... | NIFD. This dataset spans 119119119119 individuals diagnosed with different forms of FTD (FTD; age = 64.72±6.78plus-or-minus64.726.7864.72\pm 6.7864.72 ± 6.78 years, 47474747 females). In the FTD group, 52525252 individuals were diagnosed with behavioral variant FTD (BV; age = 63.07±5.77plus-or-minus63.075.7763.07\pm 5.... | ADNI. This dataset comprised of 118 individuals diagnosed with Alzheimer’s disease dementia (AD; age = 73.84±7.56plus-or-minus73.847.5673.84\pm 7.5673.84 ± 7.56 years, 56565656 females) and 206206206206 healthy individuals (HC; age = 73.87±6.39plus-or-minus73.876.3973.87\pm 6.3973.87 ± 6.39 years, 110110110110 females)... | 4RTNI. This dataset was collected as part of the 4-Repeat Tauopathy Neuroimaging Initiative (4RTNI) and used similar MRI acquisition and clinical assessments as the NIFD dataset. This dataset constituted of 59 individuals diagnosed with progressive supranuclear palsy (PSP; age = 70.79±7.65plus-or-minus70.797.6570.79\pm... | A |
Researchers have been studying that inducible defences have a major impact on predator-prey dynamics [10], e.g., size-selective predator [11]. The idea was then confirmed using a model system with inducible defences: the water flea Daphnia’s induction of a neck spine in reaction to the phantom midge Chaoborus’s predato... | The functional response is a crucial element in depicting the species’ behavioural characteristics. The dynamic behaviour of the predator is more significantly influenced by how the predator consumes their prey. Most complex dynamical behaviours, including chaotic states, periodic oscillations, stable states, etc., are... | In this article, we concentrated on a predator-prey system where the prey species choose an inducible defence strategy towards their predator. Moreover, we have chosen the Beddington-DeAngelis response, a predator-dependent functional response, where it is considered that the predator species spend some time encounteri... | This section contains the numerical validation of the dynamic behaviour of the proposed predator-prey model in the presence and absence of taxis. In the predator-prey interaction, the prey species exhibits an inducible defence strategy against their predators. Our goal is to find out the prominence or influence of this... | The Holling type II functional response was chosen by Ramos-Jiliberto et al. in their work [22] as follows: | D |
In this work, we propose an approach to modeling complex triad binding based on real biological processes. | The core of cellular immunity involves the binding of three key protein sequences: major histocompatibility complex (MHC), antigenic peptide, and T cell receptor (TCR). | In more details, the MHC first binds with the antigenic peptide to form the peptide-MHC (pMHC) complex, determining whether the peptide will be presented to the immune system. | 1 the binding of an antigenic peptide to the major histocompatibility complex (MHC), forming a peptide-MHC (pMHC) molecule (Kammertoens & Blankenstein, 2013), and | These processes involve (1) Binding of an antigenic peptide to MHC, forming a peptide-MHC (pMHC) molecule Kammertoens & Blankenstein (2013). (2) Binding of TCR to the pMHC, forming a pMHC-TCR molecule Huppa et al. (2010). | C |
These additional data demonstrate that our findings are robust and qualitatively independent of the specific choice of how dispersal is implemented. | While more complex forms of noise can be considered, the choice made here is therefore biologically relevant and convenient (Supplementary Section S1.2). | In this work, we investigate under which conditions this resistance eradication mechanism can hold in a two-dimensional metapopulation. | Extending this work to species-specific or spatially dependent migration rates would be particularly relevant for complex metapopulation structure [15, 20]. | How the environment helps shape microbial populations and species diversity [8, 9, 10, 11] is a subject of intense research [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22]. | C |
The SSI values calculated for each residue reveal those parts of the receptor that signal information about the protonation state of Asp114 by coupled conformational state changes between the rotamer states and the Asp114 protonation state (Fig. 7). | SSI highlights that changes in the protonation state of Asp114 couple to rotameric state changes on: the backbone of TM6, proximal to Asp114; the backbone of ICL2 and ICL3; residue side chains on the P-I-F, NPxxY and DRY motifs; and water sites distributed throughout the receptor. The SSI for backbone torsions is encod... | The SSI values calculated for each residue reveal those parts of the receptor that signal information about the protonation state of Asp114 by coupled conformational state changes between the rotamer states and the Asp114 protonation state (Fig. 7). | O8 is located beside Phe289 of the P-I-F motif, where we identified that backbone rotamer state changes are coupled to Asp114 protonation. | Namely, rotameric state changes in the backbone torsions of transmembrane helix TM6, proximal to Asp114, are coupled to the protonation state changes of Asp114. | D |
This paradigm was designed to be similar to the conventional oddball paradigm in which subjects were presented with a sequence of two different classes of spoken words: animal names and cardinal numbers, or color names and cardinal numbers from a loudspeaker situated one meter in front of the subject. The animal names ... | Paradigm 2 was an extension of paradigm 1, using similar sequences of twenty discrete spoken words. However, instead of a single stream of words, two competing streams were presented simultaneously by two speakers located at equal distances on either side of the subject, placed 60 degrees to the left and right (see fig... | In paradigm 3, the setup was similar to the setup of paradigm 2. The subject was presented with two competing streams from the same two speakers as in paradigm 2. However, in this case, the stimuli in each speaker were not sequences of spoken words but snippets of different stories and each snippet had a duration of ap... | To facilitate comparisons across different experimental conditions, notations AT (target event in the attended stream), AN (non-target event in the attended stream), UT (target event in the unattended stream), and UN (non-target event in the unattended stream) were used. Figure 6 shows an example of a stimulus used in ... | This paradigm was designed to be similar to the conventional oddball paradigm in which subjects were presented with a sequence of two different classes of spoken words: animal names and cardinal numbers, or color names and cardinal numbers from a loudspeaker situated one meter in front of the subject. The animal names ... | A |
Using the exonuclease frequently leads to higher accuracy (lower μ𝜇\muitalic_μ) since wrong nucleotides are preferentially discarded; but since right nucleotides are also discarded on occasion, such error correction also increases the time to copy a strand. A simple mathematical model, following numerous published mod... | Figure 2: Fast replication selects for kinetic proofreading in the presence of stalling. (a) We extend canonical models of kinetic proofreading for polymerases (i) by including the stalling effect (ii): incorporation of the correct nucleotide (here, G) at site i𝑖iitalic_i is dramatically slowed (up to 1000x [33, 34, 3... | Experiments [33, 34, 35, 36, 37, 38, 39] show that each error slows down the addition of the next base by a factor of 101−106superscript101superscript10610^{1}-10^{6}10 start_POSTSUPERSCRIPT 1 end_POSTSUPERSCRIPT - 10 start_POSTSUPERSCRIPT 6 end_POSTSUPERSCRIPT, depending on the system and the type of mismatch. We capt... | We now introduce a key experimentally observed fact — stalling [33, 34, 35, 36, 37, 38, 39, 40, 41] — that is not often considered in proofreading models, but completely inverts this intuitive trade-off picture. The incorporation of a wrong nucleotide significantly slows the catalysis of the phosphodiester bond for ... | Figure E1: Extending simple proofreading models with the stalling effect. (a) (i) A kinetic network model of proofreading at the single base level; E𝐸Eitalic_E is the enzyme (e.g., DNA polymerase loaded with a DNA template), PR,PWsubscript𝑃𝑅subscript𝑃𝑊P_{R},P_{W}italic_P start_POSTSUBSCRIPT italic_R end_POSTSUBSCR... | C |
Optimizing disease management relies heavily on the capability to forecast disease progression both promptly and precisely. | As shown in Table VI, the pretrained CDE model displays consistent results throughout the process. During training, it achieves an RMSE of 1.076, maintaining its effectiveness in the validation phase with an RMSE of 1.054. The model’s performance remains relatively stable in the testing phase as well, with an RMSE of 1... | In the evaluation of Alzheimer’s Disease, the ADAS-13 (Alzheimer’s Disease Assessment Scale-cognitive subscale with 13 items) serves as a pivotal metric for modeling cognitive impairment. An increase in the ADAS-13 scores is indicative of a deterioration in cognitive abilities, with higher scores corresponding to a mor... | As shown in Table IV, the LSTM Fusion model demonstrates a consistent performance pattern across different phases of the process. During training, it achieves an RMSE of 0.3038, maintaining its effectiveness in the validation phase with an RMSE of 0.8191. The performance remains stable in the testing phase as well, wit... | For example, in the case of cancer, identifying the early stages of the disease and disease progression through various stages (e.g., Stage I-IV), as well as predicting patient outcomes at an early stage are helpful in choosing the appropriate interventions[1]. | D |
Our primary sources of error are differences in FI and FP definitions. These were primarily in NHANES versus the other two studies, notably in terms of weakness. It seems likely that the very different prediction curves for self-reported weakness (NHANES) vs grip-strength-based weakness (HRS and ELSA) are due to differ... | The NFPFP5 appears to be an inferior measure for inferring the physical component of frailty versus the FI, as indicated by the key physical deficits – the FPFP5. The most likely reason for this is that we have generated FIs using primarily health deficits with strong physical components that likely share an etiology w... | We performed future prediction and contemporaneous inference of the FPFP5 deficits: slow gait, weakness, low body weight/weight loss, low activity, and exhaustion; for older individuals age 60+. Inference within the cross-sectional study demonstrated that many deficits, including FP frailty, can be estimated to a very ... | The FPFP5 are key physical health deficits that emerge in individuals who live with frailty. Being able to predict these deficits is useful for characterizing and forecasting individual health trajectories. We demonstrate that knowledge of an individual’s age-related health state using the FI enhances our ability to in... | Frailty is a health state characterized by reduced physiological function across multiple systems leading to an increased risk of adverse outcomes, such as disability, institutionalization, and death [1, 2, 3]. Frailty prevalence increases exponentially with age, affecting 30%-50% of individuals by age 82.[3] In additi... | C |
This dataset is sourced from a diverse set of human wastewater samples, which were processed and sequenced using deep metagenomic (next-generation) sequencing methods. | On the left, we show the training loss over one epoch of our 1.5-trillion-base-pair pretraining dataset. On the right, we show the validation loss, computed on a held-out portion of our metagenomic dataset. | Before training, we carry out byte-pair encoding (BPE) tokenization on our dataset, tailored for these nucleic acid sequences. | In developing our metagenomic foundation model, we sought a tokenization strategy that would enable high-accuracy sequence modeling, accommodate novel nucleic acid sequences, and align with best practices in modern large language models. We opted for byte-pair encoding (BPE) as our tokenization method, as it satisfies ... | Overview of METAGENE-1 and applications. Wastewater samples are collected and undergo deep metagenomic sequencing to generate DNA and RNA sequences totaling over 1.5 trillion base pairs. These sequences are tokenized using byte-pair encoding (BPE) to create the pretraining dataset. The data is used to train METAGENE-1,... | B |
We chose to focus on only 30 CpG sites because this selection strikes an optimal balance between interpretability and predictive performance. This decision is motivated by the findings of [10], whose experiments demonstrated that even with just 30 CpG sites, it is possible to achieve strong predictive accuracy. | By focusing on smaller age ranges, we can capture CpG sites exhibiting relative easy relationships with age, thus addressing age-related variability and nonlinear relationships across the broader age span. Based on these correlation coefficients, we select the top 30 CpG sites for each age group, resulting in a tailore... | Using the top 30 features identified for each age group, we train specific models tailored to each age group. | Within each age group, we calculate Pearson correlation coefficients for DNA methylation data to identify CpG sites most strongly correlated with age. This allows us to capture localized, linear relationships within each age range, potentially identifying age-specific biomarkers. | We construct an age-specific matrix for each age group and compute Pearson correlation coefficients for each CpG site within the matrix. For each age group gisubscript𝑔𝑖g_{i}italic_g start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT : | B |
Here we propose a Hebbian plasticity model in which post-synaptic neurons update their incoming weights asynchronously. Each neuron performs an update when its activity reaches a certain threshold. Importantly, we also introduce a refractory period which prevents multiple continuous updates of the same post-synaptic ne... | Neural network models with inhibition and local Hebbian plasticity have been extensively analyzed and shown to learn factorised representations of input data, which manifest in appropriate feedforward weights or receptive fields [10, 28, 24, 8, 17]. Factorised receptive fields constitute an efficient representation of ... | Lateral inhibition was first proposed by Barlow in 1952 as a mechanism to encode sensory stimuli efficiently, the so-called redundancy reduction hypothesis [2]. Initial implementations of such mechanisms can be dated back to Grossberg (1976) [13] and Rumelhart (1985) [27] where they show that constant lateral inhibitio... | In this work, we propose a biologically plausible mechanism to learn efficient factorised representations of inputs in lateral inhibition models with time-continuous plasticity. We show that it approximates the learning of classic Hebbian/anti-Hebbian networks derived from the non-negative matrix factorization. We also... | To assess the similarity between learning dynamics, we compare the learning trajectories of both asynchronous and continuous models with the discrete model. We initialize all models with the same weights and present the same stimulus sequence, and measure the cosine similarity of each neuron’s incoming weights (see App... | B |
Prior works in this context primarily focused on global properties [22, 24], lacking the capacity to condition on the geometric conditions central to our work. For instance, Bao et al. [24] demonstrate control over molecule generation based on desired quantum properties. | In the following, we investigate linker design, a subfield of fragment-based drug design. We follow Igashov et al. [13] and decompose ligands from the ZINC dataset [74] with the MMPA algorithm [75]. Note that the ZINC dataset does not contain pocket information, and the evaluated approaches operate solely at the molecu... | Research on de novo molecule generation focused extensively on generating molecules using their chemical graph representations [7, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34]. However, these methods are limited in modelling the molecules’ conformation information and are, therefore, not ideally suited for several drug-disc... | Conditional diffusion models [17, 18, 19, 20, 21] are based on the same principle but incorporate a particular condition in their training, allowing for the controlled generation. Alternatively, classifier guidance [22, 23] relies on external models for controllable generation. | Schneuing et al. [11], resulting in 130130130130 test proteins. Per target pocket, 100100100100 ligands are generated. We evaluate the generation of ligands on models that are trained on the full-atom context of the pockets in Tab. 2 and results of models trained on the Cαsubscript𝐶𝛼C_{\alpha}italic_C start_POSTSUBSC... | B |
By using a more expressive denoising network, EDM was extended to GCDM (Morehead & Cheng, 2024), which margins across conditional and unconditional settings for the QM9 dataset (Ramakrishnan et al., 2014) and the larger GEOM-Drugs dataset (Axelrod & Gómez-Bombarelli, 2022). GCDM is a diffusion model for 3D molecules th... | Figure 4: Overview of EDM (Equivariant Diffusion Models) and its extensions for molecular generation tasks. The top box represents the foundational EDM model, which uses 3D point cloud representation with E(3) equivariance to handle molecular structures. The figure highlights the key limitations of earlier models (show... | By limiting the message-passing computations to neighboring nodes, MDM (Huang et al., 2022) outperforms EDM in building chemical bonds via atom pair distances. It points out the lack of consideration for interatomic relations in GCDM, and addresses the scalability issue by introducing the Dist-transition Block. PMDM (H... | GDSS (Jo et al., 2022) is a novel permutation equivariant one-shot diffusion model. It can generate valid molecules by capturing the node-edge relationship. CDGS (Huang et al., 2023a) incoporates discrete graph structures into a diffusion model. It is permutation equivariant and implicitly defines the permutation invar... | Taking advantage of the strong relationship between the bond types and bond lengths to guide the generation of atom positions, MolDiff (Peng et al., 2023) produces high-quality 3D molecular graphs and effectively tackles the atom-bond inconsistency problem with E(3)-equivariant diffusion model. Because it models and di... | B |
It is also important to highlight the rapid advancements in large language models (LLMs), like ChatGPT, which have recently achieved remarkable breakthroughs and drawn significant attention. LLMs have shown substantial promise in bioinformatics and drug discovery tasks, offering new possibilities in these areas. Furthe... | Looking ahead, integrating LLMs with proteomic technologies could potentially revolutionize the field by enabling the personalization of anesthesia, intensive care, and pain management for individual patients. Such integration could also complement pharmacogenetic approaches to optimizing drug therapies. | In the realm of anesthesiology, our ML platform offers a new strategy for discovering potential anesthetic agents. This innovative approach has the capacity to be broadly applied to research on various conditions that impact the nervous system. With the continuous refinement of our knowledge about the mechanisms of ane... | Protemic technology has increasingly presented a vast potential in anesthesia [8], and the use of proteomic tools to study anesthetic binding sites has offered a better understanding of the mechanisms of anesthetic action. Protein-protein interaction (PPI) networks at the proteomics level provide a systematic framework... | ML technologies process and analyze vast amounts of biomedical data, thereby enhancing the efficiency of drug design and screening [40], predicting the biological activity and pharmacological properties of new molecules, optimizing drug structures, and improving binding specificity to GABA receptors. Moreover, they can... | A |
The performance of the models were quantified by correlating features of the reconstructions with the stimuli. Specifically, features from reconstructed images and original stimuli are extracted using a pretrained AlexNet model at conv1, conv2, conv3, conv4, conv5 FC6, FC7, and FC8. Subsequently, each feature layer is ... | The introduction of the IRFA model contributes several advancements to the field. First, it demonstrates that incorporating feature-based selective attention, alongside spatial attention, into brain representations significantly enhances the quality of natural image reconstruction from evoked brain responses in the V1,... | The U-NET’s output is compared with the target using three losses (an adversarial loss, a feature loss (VGG) and an L1 loss). The adversarial loss is a discriminator that is trained in parallel of the reconstruction model, which consists of 5 convolutional layers (see Fig. 1B). The feature loss uses the full set of con... | In our investigation, we explored the impact of varying the number of dedicated feature channels within our model, specifically training with 4, 16, 32, and 64 features, to determine if increasing feature separability leads to enhanced model performance. Given the convolutional nature of the model, there is an inherent... | We systematically trained the model to reconstruct images using sets of 4, 16, 32, and 64 learnable attention maps. This approach allows us to evaluate the optimal number of features required for effective reconstruction. The quality of the reconstructions is quantitatively compared with a baseline model, based on the ... | C |
From this definition, we can figure out that although the LTM processors are unconscious, they are essential for generating consciousness. Those processors are just like neurons in the human brain, each of them is unconscious, but without them, there wouldn’t be any chance to get consciousness. What’s more, conscious a... | If the consciousness of CTM is well-designed, then it must be compatible with at least most of the existing definitions of human(or animal) consciousness. So we mapped those conscious theories into CTM’s consciousness to see if they could work properly (or to say, those human conscious theories can be appropriately exp... | Since that we attempt to find the essence of the self-consciousness of a CTM, it’s necessary to present a clear and correct definition of conscious in CTM. In this section, firstly we are going to briefly introduce the definition of CTM’s consciousness given by Blums. After that, we’ll introduce some former recognized ... | The relationship between consciousness and self-consciousness is hard to tell. We couldn’t just affirm that self-consciousness is a superior kind of consciousness or consciousness is a simplified form of self-consciousness. However, we firmly believe that they are not two irrelevant concepts because self-consciousness ... | All those mentioned above are just definitions without theoretical supporting. They might sound right, but what we are pursuing is a more credible theoretical basis of consciousness. So based on some former studies about consciousness, here we present our proof of the rationalization of those definitions. | D |
L.G. acknowledges funding by the KU Leuven Research Fund (grant number C14/23/130) and the Research-Foundation Flanders (FWO, grant number G074321N). D.R.-R. is supported by the Ministry of Universities through the “Pla de Recuperacio, Transformació i Resilència” and by the EU (NextGenerationEU), together with the Univ... | The position of these bifurcations vary as a function of the control parameters of the system a,b𝑎𝑏a,bitalic_a , italic_b and ε𝜀\varepsilonitalic_ε, leading to the emergence of different dynamical behaviors and regimes. An example of the distribution of these regimes and the bifurcations which define them is illustr... | During the preparation of this work the author(s) used ChatGPT in order to receive suggestions on grammar and phrasing. All scientific content and figures were created by the authors. After using this tool, the authors reviewed the suggestions and only edited the grammar and phrasing of parts of the text. The authors t... | As observed in Fig. 2B, the main dynamics of the HH model are also observed in the FHN model, i.e. excitability and relaxation-like oscillations. Excitability is classified as type II, characterized by the lack of a distinct threshold. Excitability and oscillatory dynamics are essential to model neuronal spiking, and a... | In this most basic form, the FHN model thus consists of two coupled, nonlinear ordinary differential equations, where the first one describes the fast evolution of the membrane voltage of a neuron (u𝑢uitalic_u), while the second one represents the slow recovery through the opening of potassium channels and the inactiv... | B |
Weak coupling (0<c<0.0170𝑐0.0170<c<0.0170 < italic_c < 0.017). In this regime, the external system continues its oscillatory behavior even as the internal pulse traverses through it. However, the oscillation period in the driven regions slows down, creating a phase disparity with the non-driven regions. This discrepan... | Comparable to the previous case, phase waves that are synchronized with the driving pulse emerge within the system (Fig. 8A). In addition, the system becomes excitable under the influence of the driving pulse and is subject to perturbations from adjacent oscillatory regions. These perturbations trigger an extended resp... | Weak coupling (0<c<0.0170𝑐0.0170<c<0.0170 < italic_c < 0.017). In this regime, the external system continues its oscillatory behavior even as the internal pulse traverses through it. However, the oscillation period in the driven regions slows down, creating a phase disparity with the non-driven regions. This discrepan... | Even amidst the disruptions caused by the slower-moving traveling pulses, the phase waves linked to the driving pulse persist in their trajectory, as highlighted by the white dashed line. After the passage of the driving pulse, the traveling pulses evolve into phase waves, marked by a distinctive phase flip at the lead... | Intermediate Coupling (0.017<c<0.2950.017𝑐0.2950.017<c<0.2950.017 < italic_c < 0.295). At this level of coupling, the external system transitions to an excitable state in response to the internal pulse. Consequently, when an oscillation from a non-driven region intersects with a driven region, it triggers the latter t... | D |
Diversity characterizes the structural similarities between the generated structures and is defined using structural clustering. | Specifically, we report the number of clusters obtained when using the TM-score as the similarity metric, normalised by the number of generated sequences. | The root mean square distance (RMSD) between two structures is computed by calculating the square root of the average of the squared distances between corresponding atoms of the structures after the optimal superposition has been found. The TM-score (Y. and J., 2005) is a normalised measure of how similar two structure... | Novelty compares the structural similarity of the generated structure with a dataset of reference. Here again, we use the TM-score as our similarity metric and consider a structure as novel if its maximum TM-score against the reference dataset is below 0.5. As is commonly done, we only report the proportion of novel sa... | Designability, or self-consistency, uses ProteinMPNN (Dauparas et al., 2022) to predict a potential sequence for the generated structure and refolds it using ESMFold (Lin et al., 2022) to report the structural similarity between the original and redesigned structure. We follow the literature and report the proportion o... | A |
An interesting property of the gradient model is its representability as local dynamics and diffusive species. No nonphysical, non-local communication is required, which means that gradient calculating is reachable within biological constraints of cellular mechanisms. While cells are unlikely to keep track of an exact ... | A future research direction is deriving the brain’s slow dynamics and learning mechanisms. Training many parameters over extended periods allows black-box models, such as neural networks representing protein networks, to act as homeostatic-control agents within cells. Promising results have emerged in simplified neuron... | Building realistic models of the brain requires estimating many unobserved parameters because the complete dynamical-system state of the brain is not fully observable. It is due to this limitation that – despite nowadays being able to construct and simulate large brain models – they behave poorly once stretched beyond ... | Limitations of the method relate to the fact that brain models extend beyond the cable equation: spike propagation with delays, stochastic models, reaction kinetics and ion dynamics, and Nernst potentials are not included in the method. These can be solved mathematically but would require modifications to the brain sim... | In this brief, we introduced gradient diffusion, a methodology that facilitates the calculation of parameter gradients for any existing, unmodified model-and-neurosimulator combination, thereby enabling support for homeostatic control. This approach allows for the efficient tuning of realistic neuron models and the imp... | A |
(iii) lysogenic conversion: phage DNA is integrated into the host DNA and results in changes of the host phenotype (Barksdale and Arden, 1974; Abedon, 2008a). | Phages infect host cells by adsorbing (attaching) to receptors on the host cell wall and then delivering the genomic content into the host cytoplasm. Phages are much smaller than bacteria and each host cell presents multiple receptors that phages can bind to, so multiple phages can adsorb to a single host cell, though ... | After infection, phages replicate inside the host cell. Once the phage particle count inside the host cell has reached a threshold, say λ𝜆\lambdaitalic_λ on average, the host lyses and releases phages back into the environment. We assume that lysing occurs once host resources are exhausted, and that the replication ra... | Phages also disrupt nutrient cycling in aquatic ecosystems, where phage induced lysis (i.e. killing the host cell by rupturing the cell wall) renders organic carbon unavailable to higher trophic levels | After infection, phages replicate inside the host cell. Once the phage particle count inside the host cell has reached a threshold, say λ𝜆\lambdaitalic_λ on average, the host lyses and releases phages back into the environment. We assume that lysing occurs once host resources are exhausted, and that the replication ra... | C |
Such a setup allows the network to have memory, which is encoded in the weights, and to solve pattern recognition tasks. Suppose the network has been trained to remember two patterns corresponding to the vectors w→1subscript→𝑤1\vec{w}_{1}over→ start_ARG italic_w end_ARG start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT and w→2s... | Since the gate has two toeholds and the output has only one, the gate-output complex has a free toehold T′superscript𝑇′T^{\prime}italic_T start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT. This toehold is complementary to the free toehold T𝑇Titalic_T of the input. As a consequence, the input binds to the gate-output comple... | To summarize, what can be achieved via this process is to convert an input signal (which is a DNA strand) into a desired output signal (which is also a DNA strand) with a certain rate. In particular, this method allows to realize so-called seesaw gates Qian and Winfree (2011a, b). In addition to input and output, also ... | A key technique in this context is toehold-mediated strand displacement Yurke et al. (2000); Zhang and Winfree (2009) (see Ref. Simmel et al. (2019) for a review). This process involves a single-stranded DNA (the input) and a double-stranded DNA, whose strands are referred to as gate and output. One strand of the doubl... | Given that the neural network consists of chemicals floating around in a test tube, it is not immediately clear why these calculations should take place in the desired order (as they do in a traditional neural network implemented in a computer). For most steps, this is ensured by the fact that a reaction can only take ... | C |
The dataset underwent subject-wise 70/30 train-test splitting. Training procedures involved cross-validations on the training set. Optimization of the parameter λ𝜆\lambdaitalic_λ, was conducted by exploring different values (1, 10, 500, 5000). Following cross-validation, the model was retrained on the entire training ... | The Beta band (12-30 Hz) in the Audio model displays mixed areas with positive correlations in the frontal cortex and motor areas, linked to active thinking, focus, and motor planning. Scattered negative correlations are present across the brain, indicating variable engagement of different regions during auditory proce... | For the Alpha band (8-12 Hz), the Audio model shows positive correlations in the occipital and parietal regions, consistent with alpha rhythms’ association with relaxed states and sensory processing. Negative correlations are prominent in the frontal cortex, suggesting active suppression of irrelevant information durin... | We evaluated the reconstructed brain responses between 0.5-30 Hz (i.e. the "complete" spectrum) as well as for individual frequency bands i.e. delta, theta, alpha, and beta (Abhang et al., 2016). Delta frequencies are typically between 0.5 and 4 Hz, often associated with deep sleep or states of unconsciousness. Theta f... | Figure 2: Pearson Correlation topography maps (subject-wise average), visualizing the performance of all encoding strategies (TFD, wav2vec2, CLIP, GPT-2) performance for every sensor and frequency band, as well as the full spectrum case ("complete"). In the case of audio encoders, high values of correlation occur in la... | C |
We consider two different types of attractant concentration profiles. In one case all the concentration peaks are of the same height, and in the other case they have unequal heights. We use a sine function, and superposition of two sine functions to represent these two cases (also see Fig. 1) but our conclusions do not... | In this section, we present our results for the case when the attractant concentration shows a sinusoidal variation in space, as shown in Fig. 1 (solid line). In Fig. 3 we plot the position distribution P(x)𝑃𝑥P(x)italic_P ( italic_x ) of the agent in the long time limit. We find that RL algorithm can successfully ge... | We consider two different types of attractant concentration profiles. In one case all the concentration peaks are of the same height, and in the other case they have unequal heights. We use a sine function, and superposition of two sine functions to represent these two cases (also see Fig. 1) but our conclusions do not... | In Sec. II we describe the model in details and explain the RL algorithm. In Sec. III we present our results for the sine wave attractant profile and in Sec. IV we take up superposition of two sine waves. We include some concluding remarks in Sec. V. | We consider a one dimensional system of size L𝐿Litalic_L with periodic boundary condition across which a spatially varying attractant concentration profile [L](x)delimited-[]𝐿𝑥[L](x)[ italic_L ] ( italic_x ) is set up. A successful RL strategy should allow the agent to efficiently localize near the maxima of [L](x... | C |
In our model, each neural response is associated with a specific spatial location within the visual field. Adapting the model to accommodate multiple receptive fields per neural response could significantly enhance its applicability, particularly because recording sites often contain signals from multiple neurons. Addi... | Beyond enhancing our understanding of visual processes, neural encoding models also hold potential for applied domains. For instance, these models can facilitate advancements in cortical prosthetics, potentially improving the accuracy of prosthetic virtual reality simulations that aim to stimulate visual perceptions wi... | One common strategy in neural encoding involves leveraging nonlinear features extracted from deep neural networks trained on categorization tasks. These features are used to encode neural responses in visual areas by linearly mapping the extracted visual features to observed neural activity [2, 3, 4, 5, 6, 7]. While th... | Neural encoding models, particularly those designed to predict neural activity from naturalistic images, significantly enhance our understanding of how visual stimuli are processed and represented in the brain. These models, incorporating aspects of retinotopy, are pivotal in elucidating the complex relationship betwee... | In our model, each neural response is associated with a specific spatial location within the visual field. Adapting the model to accommodate multiple receptive fields per neural response could significantly enhance its applicability, particularly because recording sites often contain signals from multiple neurons. Addi... | A |
Entering into gene details, both methods were able to identify common genes for each cluster studied (Pde10a, Cdh18, Penk Sfmbt2 for cluster iSPN; Pde10a, Cdh18, Fam155a, Rgs9 for cluster dSPN). From these genes, we identify new HD-related genes that are common to both methods: Cdh18 on neuronal SPNs clusters (SAHP: iS... | Interestingly, however, we find genes that are only identified when using the proposed SHAP analysis. Some of these have been previously described in the context of HD, including Rarb [15], Cntn6 genes[16] or Onecut2 [17]. Importantly, we also identified informative genes that have not been previously described in the ... | Commonly used DGE techniques such as DESeq2 provide useful insights into gene expression changes between conditions. However, these techniques are not able to capture the association and interaction of genes at different levels. In contrast, the XAI-based approach adopted in this paper is able to analyze how individual... | GSEA results from both DESeq2 and SHAP are shown in Table II. Common to both methods, we find a decreased synaptic function in SPN clusters. However, GSEA from DESeq2 data emphasizes the decrease in a global broader list of synaptic-related regulators included in the category named Synapse, whereas GSEA from SHAP manif... | Further model validation was made from SHAP results based on published data obtained from in vivo HD models or patients. For this validation, we compare the correlation between SHAP values and gene expression of previously described altered genes, expecting to observe a positive correlation between higher HD expression... | A |
An under-explored area is the incorporation of short indels, which can potentially be treated as another type of count-based variation. | Note that not all data types are well suited for the Pool-seq approach. For instance, a widespread practice is to use a variant calling tool on the data before downstream analyses. However, many standard variant callers were developed for individual instead of pooled data, meaning that their statistical assumptions mig... | In the future, given the ease with which statistics computations can be incorporated into our modular software design, we aim to re-implement more of the existing Pool-seq statistics, such as f𝑓fitalic_f statistics (5, 6), and implement a Pool-seq-based GWA tool (1). | Furthermore, we want to integrate grenedalf with our short-read processing and variant calling pipeline grenepipe (17), which already supports estimating allele frequencies from Pool-seq data via the HAF-pipe tool (7, 8). | Most commonly, our input are sequence reads or read-derived allele counts, as those fully capture the effects of both sources of noise, which can then be corrected for. Our implementation however can also be used with inferred or adjusted allele frequencies as input, for instance using information from the haplotype fr... | C |
Matching: Matching reflects the precision of the model by quantifying how many generated conformations closely resemble those in the reference set within a defined RMSD threshold. | Coverage: This metric measures the diversity of the generated conformations by indicating the proportion of reference conformations that are matched within a specified RMSD threshold. | Minimum RMSD: This metric provides insight into the average best-case alignment between the generated conformations and the reference set, indicating the overall accuracy of the model. | Our primary objective is to enhance the generation of 3D molecular structures in a way that closely aligns them with the actual models, regardless of their orientation or positioning. To quantify the precision of this alignment, we utilize the post-alignment Root-Mean-Square Deviation (RMSD), a commonly used metric in ... | MAT-P (Mean RMSD-Precision): MAT-P scores reflect the mean RMSD between each generated conformation and its nearest reference counterpart. It calculates the average structural deviation between the generated and reference conformations. Low MAT-P scores indicate that the generated conformations closely resemble the ref... | B |
In Appendix B we’ll see how the results (24) - (27) can be generalized in the case when advection is taken into account in addition to diffusion. | The first impression is that the problem is overdetermined, and the existence of the solution of (10) demands | The connection between the existence of the elementary solution of (10) and the possible symmetry of the equation will be analysed in our next publication. | We did finetune the speed c𝑐citalic_c in the previous Section for the solution of (21) to be expressed in terms of elementary functions, and not for the solution to exist. | In this paper we will use the method of exclusion of the independent variable benguria ; rosu ; kogan3 ; kogan4 ; kogan5 , which allows to reduce the problem of integration of the second order ODE to the problem of integration of the first order ODE, to obtain the exact solutions of Eq. (6). Thus we were able to obtain... | B |
Seed Production: Adult individuals of each species produce seeds, with the quantity determined by their reproductive strategies and environmental conditions. | The life-history difference between the two species is expressed in two parameters: the probability of a juvenile reaching maturity and the number of seeds produced by each adult. In particular, the probability of each juvenile to mature and produce seeds is pfsubscript𝑝𝑓p_{f}italic_p start_POSTSUBSCRIPT italic_f end... | Adult Mortality: At the end of each generation, all adult individuals die, ensuring complete generational turnover and the absence of overlap between successive generations. | Maturation and Seed Production: Each juvenile has a species-specific probability of maturing into an adult capable of producing seeds, thus forming the next generation. | Seed Production: Adult individuals of each species produce seeds, with the quantity determined by their reproductive strategies and environmental conditions. | B |
This model is able to track the infections patterns among age groups happening in different exposition contexts. In this work, we exploited this model to reconstruct COVID-19 spread in different scenarios between September 2020 and July 2021 in Italy, in Lombardy and in Lazio, and taking into account the vaccination ca... | This model allows to assess the impact of different NPIs that can be implemented in various contexts heterogeneously. | In this way, we can interpret the numerical results in light of the specific NPIs placed in act and evaluate the impact of specific interventions. | In this paper, we retrospectively assess the effectiveness of restrictions associated with the introduction of the Green Pass, an NPI implemented in Italy in 2021 to maximize vaccination coverage, particularly in public exposure contexts. By incorporating the Green Pass into the model, we aim to simulate a realistic sc... | Finally, we conclude that the Green Pass NPIs impact more in the context of exposition of work, and in a lighter way in leisure, as we can see in Figure 25. On the other hand, an increase in the rate of exposure to the virus is observed in contexts such as school and home. | A |
In this work, we have introduced new topological descriptors (PSG) and showed their value in clustering the library of RNA motifs trained by existing RNAs to predict RNA-like hypothetical topologies. We showed that the k𝑘kitalic_k-means clustering method performs especially well among all six clustering methods consid... | For each graph, we can calculate its persistent betti 0 (β00,dsuperscriptsubscript𝛽00𝑑\beta_{0}^{0,d}italic_β start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 0 , italic_d end_POSTSUPERSCRIPT) and persistent betti 1 (β10,dsuperscriptsubscript𝛽10𝑑\beta_{1}^{0,d}italic_β start_POSTSUBSCRIPT 1 end_POSTSUB... | 2) RNA-like graphs do not have Betti 1 bars, while non-RNA-like graphs are more likely to have at least one Betti 1 bar. The same patterns exist in all graphs illustrated in Figure 4. Details can be found in the Supporting Information. Therefore, we suggest that future research can follow these rules in designing new R... | In this work, we have introduced new topological descriptors (PSG) and showed their value in clustering the library of RNA motifs trained by existing RNAs to predict RNA-like hypothetical topologies. We showed that the k𝑘kitalic_k-means clustering method performs especially well among all six clustering methods consid... | Our work indicates that the RNA-like universe is at least 46% and that biological RNA topologies are more likely to contain subgraphs that distinguish them from non-RNA-like structures. Echoing this finding from the perspective of Persistent Spectral Graph (PSG) analysis is that RNA-like graphs tend to exhibit more cha... | D |
Eukaryotic cells are organized into functional compartments, such as the nucleus and the cytoplasm. The former houses the genome and DNA-related machinery, the latter contains cytosol and organelles for cellular functions such as protein synthesis and degradation [1]. The boundary between the nucleus and the cytoplasm ... | While individual molecular translocations through NPCs do not require energy, being thermally driven and facilitated by interactions with nucleoporins inside the NPC [9, 10, 11, 12], they are part of a complex cycle that is essentially an energy-driven pump [12, 13, 14, 15, 16]. This cycle can generate import/export fl... | The import–export cycles of cargo proteins are tightly regulated by RanGTP. Figure 1 depicts the import cycle, with each cycle using one GTP molecule and resulting in the export of one Ran molecule per cargo 111For transport proteins like Importin-β𝛽\betaitalic_β, which utilize an adaptor protein, Importin-α𝛼\alphait... | where the last term is the exchange of RanGTP–NTR through the NPCs, with 𝒱csubscript𝒱𝑐\mathcal{V}_{c}caligraphic_V start_POSTSUBSCRIPT italic_c end_POSTSUBSCRIPT as the volume of cytoplasm, whereas the first is the loss due to the GTP hydrolysis by RanGAP, converting RanGTP to RanGDP at a rate η𝜂\etaitalic_η. Simil... | To date, this system has been primarily modeled using kinetic rate equations [25, 26, 27, 28, 29, 30, 31, 32, 33, 34], while assuming a rapid homogenization of the Ran cycle components within the nucleus and cytoplasm [35, 36, 3, 31]. However, this assumption breaks down if the sources, or sinks, of the molecular compo... | A |
The authors are indebted to Marco Giulini for technical support and an insightful reading of the manuscript. | The aforementioned studies described in detail the structural organisation of CzrA and provided a convincing picture of the local and global changes that occur upon binding with the DNA and/or the zinc ions; in particular, the computational investigation carried out by Chakravorty and coworkers [12] has demonstrated th... | RP acknowledges support from ICSC - Centro Nazionale di Ricerca in HPC, Big Data and Quantum Computing, funded by the European Union under NextGenerationEU. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or The European Research Executive ... | Even more compelling is the MEOW analysis of the DNA binding region of CzrA, for which, opposite to what was observed in the case of the zinc coordination site, the results highlight an increment in the relevance of the associated residues when the molecule is bound to the metal ions. The MEOW information fields of the... | The next natural step in the analysis is to investigate how the zinc coordination state impacts the protein’s fluctuation patterns around such conformations. The structural variability of the two forms of CzrA is thus inspected on a residue basis in Fig. 3, where we compare the apo and holo systems’ root mean square fl... | B |
In other words, λ~μsubscript~𝜆𝜇\tilde{\lambda}_{\mu}over~ start_ARG italic_λ end_ARG start_POSTSUBSCRIPT italic_μ end_POSTSUBSCRIPT is determined by {Θ,θμ;Λμ(t)}Θsubscript𝜃𝜇subscriptΛ𝜇𝑡\left\{\Theta,\theta_{\mu};\Lambda_{\mu}(t)\right\}{ roman_Θ , italic_θ start_POSTSUBSCRIPT italic_μ end_POSTSUBSCRIPT ; roman_Λ... | To use the Lindstrom-Bates empirical Bayes (LBEB) approach for NLME [28], we need to make some statistical assumptions: | The only way to achieve a true patient specific model of that patient’s disease state is through experimentation, i.e. administering brief doses of Abiraterone and measuring PSA levels over time to fit all the parameters to the measured PSA trajectories for each patient. Once parameters are established, the oncologist ... | The model is nonlinear in both common as well as person specific (mixed) factors. For this reason, we utilize an established statistical approach, the empirical Bayes approach of Lindstrom and Bates [28], to fit the model to the normalized PSA data within the framework of nonlinear mixed effects (NLME). The estimated p... | In this work, we use a systematic statistical method, LBEB approach for NLME, to estimate patient-informed parameters for a Stackelberg game-theoretic multi-population model that quantitatively describes prostate cancer progression [22]. We obtain estimates for common parameters across all patients as well as the proba... | A |
The properties of this function are determined by the covariance function k(𝐱,𝐱′)𝑘𝐱superscript𝐱′k\left(\mathbf{x},\mathbf{x^{\prime}}\right)italic_k ( bold_x , bold_x start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT ), which measures the similarity between two inputs 𝐱𝐱\mathbf{x}bold_x and 𝐱′superscript𝐱′\mathbf{x... | The atomistic covariance functionkatomistic(xj,xl′)subscript𝑘atomisticsubscriptx𝑗subscriptsuperscriptx′𝑙k_{\text{atomistic}}\left(\textbf{x}_{j},\textbf{x}^{\prime}_{l}\right)italic_k start_POSTSUBSCRIPT atomistic end_POSTSUBSCRIPT ( x start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT , x start_POSTSUPERSCRIPT ′ end_P... | The GPR framework provides a predictive distribution for the output corresponding to any input x. The actually predicted value for the output of an input 𝐱𝐱\mathbf{x}bold_x is the mean of this predictive distribution, given as | As a local representation, we used Smooth overlap of atomic positions (SOAP) [25]. It is commonly used and GPR models with SOAP representations as inputs have shown high accuracy [29]. For atomistic representations, an input x will consist of representations xjsubscriptx𝑗\textbf{x}_{j}x start_POSTSUBSCRIPT italic_j en... | In our case, an input x to the model can be either a global representation of the atomistic system or local representations of the atomistic environments. The corresponding output y𝑦yitalic_y is a scalar energy value. | D |
Our results show that XATGRN consistently outperforms state-of-the-art models across multiple datasets. The cross-attention mechanism allows XATGRN to focus on the most relevant features from bulk gene expression data, while our relation graph embedding module effectively captures both connectivity and directionality w... | Extensive experiments on benchmark datasets underscore the model’s effectiveness in uncovering previously unknown regulatory mechanisms and its potential to identify novel therapeutic targets for complex diseases. Our XATGRN model provides a comprehensive and powerful framework for advancing our understanding of gene r... | We introduce the XATGRN model, which is capable of predicting the existence, directionality, and type of regulatory relationships in Gene Regulatory Networks (GRNs). This model offers a comprehensive understanding of the intricate mechanisms of gene regulation. | We conduct extensive experiments on multiple benchmark datasets, demonstrating XATGRN’s proficiency in uncovering unseen regulatory mechanisms and potential therapeutic targets for complex diseases. | In conclusion, our XATGRN represents a significant step forward in GRN inference, providing a robust and accurate framework for studying gene regulatory mechanisms. By effectively managing skewed degree distributions and leveraging advanced attention mechanisms, XATGRN serves as a powerful tool for uncovering regulator... | A |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 4