Spatially resolved transcriptomics (SRT) allows for the comprehensive profiling of gene expression while preserving spatial context, advancing the study of tissue architecture. However, existing computational approaches still face key limitations, particularly the insufficient exploitation of histology information and the lack of cross-modal meaningful contrastive strategies for biological analyses. To overcome these challenges, we propose GCAST, a graph contrastive autoencoder framework for spatial transcriptomics that seamlessly integrates multimodal SRT data. GCAST adopts a self-supervised strategy to derive biologically meaningful representations directly from histology images when available. GCAST constructs dual graph views based on data augmentation and introduces a novel contrastive learning designed to leverage histology-weighted and gene-weighted features and improve biological interpretability. In addition, GCAST employs a block-diagonal graph construction to automatically align multiple datasets, achieving batch-effect correction without manual intervention. The framework not only captures spatial gene expression patterns to identify tissue domains but also adapts to datasets with or without histological images and supports the integration of multiple datasets for joint analyses. Overall, GCAST provides a unified and biologically informed framework that has the potential to facilitate deeper analyses of spatial transcriptomics.

Accurately predicting the pathogenicity of missense variants is crucial for improving disease diagnosis and advancing clinical research. However, existing computational methods primarily focus on general pathogenicity predictions, overlooking assessments of disease-specific conditions. In this study, we propose DS-MVP, a method capable of predicting disease-specific pathogenicity of missense variants in human genomes. DS-MVP first leverages a deep learning model pre-trained on a large general pathogenicity dataset to learn rich representation of missense variants. It then fine-tunes these representations with an XGBoost model on smaller datasets for specific diseases. We evaluated the learned representation by testing it on multiple binary pathogenicity datasets and gene-level statistics, demonstrating that DS-MVP outperforms existing state-of-the-art methods, such as MetaRNN and AlphaMissense. Additionally, DS-MVP excels in multi-label and multi-class classification, effectively classifying disease-specific pathogenic missense variants based on disease conditions. It further enhances predictions by fine-tuning the pre-trained model on disease-specific datasets. Finally, we analyzed the contributions of the pre-trained model and various feature types, with gene description corpus features from large language model and genetic feature fusion contributing the most. These results underscore that DS-MVP represents a broader perspective on pathogenicity prediction and holds potential as an effective tool for disease diagnosis.

MOTIVATION: Binding of peptides to major histocompatibility complex (MHC) molecules plays a crucial role in triggering T cell recognition mechanisms essential for immune response. Accurate prediction of MHC-peptide binding is vital for the development of cancer therapeutic vaccines. While recent deep learning-based methods have achieved significant performance in predicting MHC-peptide binding affinity, most of them separately encode MHC molecules and peptides as inputs, potentially overlooking critical interaction information between the two. RESULTS: In this work, we propose RPEMHC, a new deep learning approach based on residue-residue pair encoding to predict the binding affinity between peptides and MHC, which encode an MHC molecule and a peptide as a residue-residue pair map. We evaluate the performance of RPEMHC on various MHC-II-related datasets for MHC-peptide binding prediction, demonstrating that RPEMHC achieves better or comparable performance against other state-of-the-art baselines. Moreover, we further construct experiments on MHC-I-related datasets, and experimental results demonstrate that our method can work on both two MHC classes. These extensive validations have manifested that RPEMHC is an effective tool for studying MHC-peptide interactions and can potentially facilitate the vaccine development. AVAILABILITY: The source code of the method along with trained models is freely available at https://github.com/lennylv/RPEMHC.

This study aims to tackle the intricate challenge of predicting RNA-small molecule binding sites to explore the potential value in the field of RNA drug targets. To address this challenge, we propose the MultiModRLBP method, which integrates multi-modal features using deep learning algorithms. These features include 3D structural properties at the nucleotide base level of the RNA molecule, relational graphs based on overall RNA structure, and rich RNA semantic information. In our investigation, we gathered 851 interactions between RNA and small molecule ligand from the RNAglib dataset and RLBind training set. Unlike conventional training sets, this collection broadened its scope by including RNA complexes that have the same RNA sequence but change their respective binding sites due to structural differences or the presence of different ligands. This enhancement enables the MultiModRLBP model to more accurately capture subtle changes at the structural level, ultimately improving its ability to discern nuances among similar RNA conformations. Furthermore, we evaluated MultiModRLBP on two classic test sets, Test18 and Test3, highlighting its performance disparities on small molecules based on metal and non-metal ions. Additionally, we conducted a structural sensitivity analysis on specific complex categories, considering RNA instances with varying degrees of structural changes and whether they share the same ligands. The research results indicate that MultiModRLBP outperforms the current state-of-the-art methods on multiple classic test sets, particularly excelling in predicting binding sites for non-metal ions and instances where the binding sites are widely distributed along the sequence. MultiModRLBP also can be used as a potential tool when the RNA structure is perturbed or the RNA experimental tertiary structure is not available. Most importantly, MultiModRLBP exhibits the capability to distinguish binding characteristics of RNA that are structurally diverse yet exhibit sequence similarity. These advancements hold promise in reducing the costs associated with the development of RNA-targeted drugs.

MOTIVATION: Accurate and rapid prediction of protein-ligand binding affinity is a great challenge currently encountered in drug discovery. Recent advances have manifested a promising alternative in applying deep learning-based computational approaches for accurately quantifying binding affinity. The structure complementarity between protein-binding pocket and ligand has a great effect on the binding strength between a protein and a ligand, but most of existing deep learning approaches usually extracted the features of pocket and ligand by these two detached modules. RESULTS: In this work, a new deep learning approach based on the cross-attention mechanism named CAPLA was developed for improved prediction of protein-ligand binding affinity by learning features from sequence-level information of both protein and ligand. Specifically, CAPLA employs the cross-attention mechanism to capture the mutual effect of protein-binding pocket and ligand. We evaluated the performance of our proposed CAPLA on comprehensive benchmarking experiments on binding affinity prediction, demonstrating the superior performance of CAPLA over state-of-the-art baseline approaches. Moreover, we provided the interpretability for CAPLA to uncover critical functional residues that contribute most to the binding affinity through the analysis of the attention scores generated by the cross-attention mechanism. Consequently, these results indicate that CAPLA is an effective approach for binding affinity prediction and may contribute to useful help for further consequent applications. AVAILABILITY AND IMPLEMENTATION: The source code of the method along with trained models is freely available at https://github.com/lennylv/CAPLA. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Accurate identification of RNA modification sites is of great significance in understanding the functions and regulatory mechanisms of RNAs. Recent advances have shown great promise in applying computational methods based on deep learning for accurate prediction of RNA modifications. However, those methods generally predicted only a single type of RNA modification. In addition, such methods suffered from the scarcity of the interpretability for their predicted results. In this work, a new Transformer-based deep learning method was proposed to predict multiple RNA modifications simultaneously, referred to as TransRNAm. More specifically, TransRNAm employs Transformer to extract contextual feature and convolutional neural networks to further learn high-latent feature representations of RNA sequences relevant for RNA modifications. Importantly, by integrating the self-attention mechanism in Transformer with convolutional neural network, TransRNAm is capable of not only capturing the critical nucleotide sites that contribute significantly to RNA modification prediction, but also revealing the underlying association among different types of RNA modifications. Consequently, this work provided an accurate and interpretable predictor for multiple RNA modification prediction, which may contribute to uncovering the sequence-based forming mechanism of RNA modification sites.

Effectively and accurately predicting the effects of interactions between proteins after amino acid mutations is a key issue for understanding the mechanism of protein function and drug design. In this study, we present a deep graph convolution (DGC) network-based framework, DGCddG, to predict the changes of protein-protein binding affinity after mutation. DGCddG incorporates multi-layer graph convolution to extract a deep, contextualized representation for each residue of the protein complex structure. The mined channels of the mutation sites by DGC is then fitted to the binding affinity with a multi-layer perceptron. Experiments with results on multiple datasets show that our model can achieve relatively good performance for both single and multi-point mutations. For blind tests on datasets related to angiotensin-converting enzyme 2 binding with the SARS-CoV-2 virus, our method shows better results in predicting ACE2 changes, may help in finding favorable antibodies. Code and data availability: https://github.com/lennylv/DGCddG.

Proteinprotein interactions are the basis of many cellular biological processes, such as cellular organization, signal transduction, and immune response. Identifying proteinprotein interaction sites is essential for understanding the mechanisms of various biological processes, disease development, and drug design. However, it remains a challenging task to make accurate predictions, as the small amount of training data and severe imbalanced classification reduce the performance of computational methods. We design a deep learning method named ctP2ISP to improve the prediction of proteinprotein interaction sites. ctP2ISP employs Convolution and Transformer to extract information and enhance information perception so that semantic features can be mined to identify proteinprotein interaction sites. A weighting loss function with different sample weights is designed to suppress the preference of the model toward multi-category prediction. To efficiently reuse the information in the training set, a preprocessing of data augmentation with an improved sample-oriented sampling strategy is applied. The trained ctP2ISP was evaluated against current state-of-the-art methods on six public datasets. The results show that ctP2ISP outperforms all other competing methods on the balance metrics: F1, MCC, and AUPRC. In particular, our prediction on open tests related to viruses may also be consistent with biological insights.

The binding of DNA sequences to cell type-specific transcription factors is essential for regulating gene expression in all organisms. Many variants occurring in these binding regions play crucial roles in human disease by disrupting the cis-regulation of gene expression. We first implemented a sequence-based deep learning model called deepBICS to quantify the intensity of transcription factors-DNA binding. The experimental results not only showed the superiority of deepBICS on ChIP-seq data sets but also suggested deepBICS as a language model could help the classification of disease-related and neutral variants. We then built a language model-based method called deepBICS4SNV to predict the pathogenicity of single nucleotide variants. The good performance of deepBICS4SNV on 2 tests related to Mendelian disorders and viral diseases shows the sequence contextual information derived from language models can improve prediction accuracy and generalization capability.

Spiking neural P systems with communication on request (SNQP systems) are neurally inspired computing devices, where a neuron actively seeks spikes from presynaptic neurons instead of passively waiting for spikes. In this work, we consider SNQP systems with mute rules (SNQPM systems), where mute rules have no communication functioning, namely the application of a mute rule only affects the number of spikes in the neuron where the rule resides, without effect on other neurons. It is demonstrated the computation capability of SNQPM systems with only mute rules does not exceed that of register machines with two registers, thereby not Turing universal. SNQPM systems are Turing universal when both mute rules and request rules are employed. Furthermore, two universal SNQPM systems with 7 neurons or 13 neurons are constructed as devices of number generating and function computing, respectively. Comparing to the universal SNQP system with 14 neurons and two types of spikes, SNQPM systems show the capability of trading-off mute rules and the types of spikes.

MOTIVATION: Histone modifications are epigenetic markers that impact gene expression by altering the chromatin structure or recruiting histone modifiers. Their accurate identification is key to unraveling the mechanisms by which they regulate gene expression. However, the solutions for this task can be improved by exploiting multiple relationships from dataset and exploring designs of learning models, for example jointly learning technology. RESULTS: This article proposes a deep learning-based multi-objective computational approach, iHMnBS, to identify which of the seven typical histone modifications a DNA sequence may choose to bind, and which parts of the DNA sequence bind to them. iHMnBS employs a customized dataset that allows the marking of modifications contained in histones that may bind to any position in the DNA sequence. iHMnBS tries to mine the information implicit in this richer data by means of deep neural networks. In comprehensive comparisons, iHMnBS outperforms a baseline method, and the probability of binding to modified histones assigned to a representative nucleotide of a DNA sequence can serve as a reference for biological experiments. Since the interaction between transcription factors and histone modifications has an important role in gene expression, we extracted a number of sequence patterns that may bind to transcription factors, and explored their possible impact on disease. AVAILABILITY AND IMPLEMENTATION: The source code is available at https://github.com/lennylv/iHMnBS. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Protein structure can be severely disrupted by frameshift and non-sense mutations at specific positions in the protein sequence. Frameshift and non-sense mutation cases can also be found in healthy individuals. A method to distinguish neutral and potentially disease-associated frameshift and non-sense mutations is of practical and fundamental importance. It would allow researchers to rapidly screen out the potentially pathogenic sites from a large number of mutated genes and then use these sites as drug targets to speed up diagnosis and improve access to treatment. The problem of how to distinguish between neutral and potentially disease-associated frameshift and non-sense mutations remains under-researched. RESULTS: We built a Transformer-based neural network model to predict the pathogenicity of frameshift and non-sense mutations on protein features and named it TransPPMP. The feature matrix of contextual sequences computed by the ESM pre-training model, type of mutation residue and the auxiliary features, including structure and function information, are combined as input features, and the focal loss function is designed to solve the sample imbalance problem during the training. In 10-fold cross-validation and independent blind test set, TransPPMP showed good robust performance and absolute advantages in all evaluation metrics compared with four other advanced methods, namely, ENTPRISE-X, VEST-indel, DDIG-in and CADD. In addition, we demonstrate the usefulness of the multi-head attention mechanism in Transformer to predict the pathogenicity of mutations-not only can multiple self-attention heads learn local and global interactions but also functional sites with a large influence on the mutated residue can be captured by attention focus. These could offer useful clues to study the pathogenicity mechanism of human complex diseases for which traditional machine learning methods fall short. AVAILABILITY AND IMPLEMENTATION: TransPPMP is available at https://github.com/lennylv/TransPPMP. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Spiking neural P systems (abbreviated as SNP systems) are models of computation that mimic the behavior of biological neurons. The spiking neural P systems with communication on request (abbreviated as SNQP systems) are a recently developed class of SNP system, where a neuron actively requests spikes from the neighboring neurons instead of passively receiving spikes. It is already known that small SNQP systems, with four unbounded neurons, can achieve Turing universality. In this context, 'unbounded' means that the number of spikes in a neuron is not capped. This work investigates the dependency of the number of unbounded neurons on the computation capability of SNQP systems. Specifically, we prove that (1) SNQP systems composed entirely of bounded neurons can characterize the family of finite sets of numbers; (2) SNQP systems containing two unbounded neurons are capable of generating the family of semilinear sets of numbers; (3) SNQP systems containing three unbounded neurons are capable of generating nonsemilinear sets of numbers. Moreover, it is obtained in a constructive way that SNQP systems with two unbounded neurons compute the operations of Boolean logic gates, i.e., OR, AND, NOT, and XOR gates. These theoretical findings demonstrate that the number of unbounded neurons is a key parameter that influences the computation capability of SNQP systems.

Spiking neural P systems (SNP systems) are a class of distributed and parallel computation models, which are inspired by the way in which neurons process information through spikes, where the integrate-and-fire behavior of neurons and the distribution of produced spikes are achieved by spiking rules. In this work, a novel mechanism for separately describing the integrate-and-fire behavior of neurons and the distribution of produced spikes, and a novel variant of the SNP systems, named evolution-communication SNP (ECSNP) systems, is proposed. More precisely, the integrate-and-fire behavior of neurons is achieved by spike-evolution rules, and the distribution of produced spikes is achieved by spike-communication rules. Then, the computational power of ECSNP systems is examined. It is demonstrated that ECSNP systems are Turing universal as number-generating devices. Furthermore, the computational power of ECSNP systems with a restricted form, i.e. the quantity of spikes in each neuron throughout a computation does not exceed some constant, is also investigated, and it is shown that such restricted ECSNP systems can only characterize the family of semilinear number sets. These results manifest that the capacity of neurons for information storage (i.e. the quantity of spikes) has a critical impact on the ECSNP systems to achieve a desired computational power.

Spiking neural P (SN P) systems are a class of discrete neuron-inspired computation models, where information is encoded by the numbers of spikes in neurons and the timing of spikes. However, due to the discontinuous nature of the integrate-and-fire behavior of neurons and the symbolic representation of information, SN P systems are incompatible with the gradient descent-based training algorithms, such as the backpropagation algorithm, and lack the capability of processing the numerical representation of information. In this work, motivated by the numerical nature of numerical P (NP) systems in the area of membrane computing, a novel class of SN P systems is proposed, called numerical SN P (NSN P) systems. More precisely, information is encoded by the values of variables, and the integrate-and-fire way of neurons and the distribution of produced values are described by continuous production functions. The computation power of NSN P systems is investigated. We prove that NSN P is Turing universal as number generating devices, where the production functions in each neuron are linear functions, each involving at most one variable; as number accepting devices, NSN P systems are proved to be universal as well, even if each neuron contains only one production function. These results show that even if a single neuron is simple in the sense that it contains one or two production functions and the production functions in each neuron are linear functions with one variable, a network of simple neurons are still computationally powerful. With the powerful computation power and the characteristic of continuous production functions, developing learning algorithms for NSN P systems is potentially exploitable.

Spiking neural P systems (SN P systems) are a class of parallel computing models, inspired by the way in which neurons process information and communicate to each other by means of spikes. In this work, we consider a variant of SN P systems, SN P systems with polarizations (PSN P systems), where the integrate-and-fire conditions are associated with polarizations of neurons. The computation power of PSN P systems working in the asynchronous mode (at a computation step, a neuron with enabled rules does not obligatorily fire), instead of the synchronous mode (a neuron with enabled rules should fire), is investigated. We proved that asynchronous PSN P systems with extended rules (the application of a rule can produce more than one spikes) or standard rules (all rules can only produce a spike) can both characterize partially blind counter machines, hence, such systems are not Turing universal. The equivalence of the computation power of asynchronous PSN P systems in both cases of using extended rules or standard rules indicates that asynchronous PSN P systems are robust in terms of the amount of information exchanged among neurons. It is known that synchronous PSN P systems with standard rules are Turing universal, so these results also suggest that the working model, synchronization or asynchronization, is an essential ingredient for a PSN P system to achieve a powerful computation capability.

Spiking neural P systems are a class of third generation neural networks belonging to the framework of membrane computing. Spiking neural P systems with communication on request (SNQ P systems) are a type of spiking neural P system where the spikes are requested from neighboring neurons. SNQ P systems have previously been proved to be universal (computationally equivalent to Turing machines) when two types of spikes are considered. This paper studies a simplified version of SNQ P systems, i.e. SNQ P systems with one type of spike. It is proved that one type of spike is enough to guarantee the Turing universality of SNQ P systems. Theoretical results are shown in the cases of the SNQ P system used in both generating and accepting modes. Furthermore, the influence of the number of unbounded neurons (the number of spikes in a neuron is not bounded) on the computation power of SNQ P systems with one type of spike is investigated. It is found that SNQ P systems functioning as number generating devices with one type of spike and four unbounded neurons are Turing universal.

Spiking neural P (SN P) systems are a class of parallel computation models inspired by neurons, where the firing condition of a neuron is described by a regular expression associated with spiking rules. However, it is NP-complete to decide whether the number of spikes is in the length set of the language associated with the regular expression. In this paper, in order to avoid using regular expressions, two major and rather natural modifications in their form and functioning are proposed: the spiking rules no longer check the number of spikes in a neuron, but, in exchange, a polarization is associated with neurons and rules, one of the three electrical charges -, 0,+. Surprisingly enough, the computing devices obtained are still computationally complete, which are able to compute all Turing computable sets of natural numbers. On this basis, the number of neurons in a universal SN P system with polarizations is estimated. Several research directions are mentioned at the end of this paper.

Spiking neural P systems with rules on synapses (RSSN P systems, for short) are a class of distributed and parallel computation models inspired by the way in which neurons process and communicate information with each other by means of spikes, where neurons only contain spikes and the evolution rules are on synapses. RSSN P systems have been proved to be Turing universal, using the strategy that restricts all the applied rules to consume the same number of spikes from the given neuron, termed as equal spikes consumption strategy. In this work, in order to avoid imposing the equal spikes consumption restriction on the application of rules, a new strategy for rule application, termed as sum spikes consumption strategy, is considered in RSSN P systems, where a maximal set of enabled rules from synapses starting from the same neuron is nondeterministically chosen to be applied, in the sense that no further synapse can use any of its rules, and the sum of these numbers of spikes that all the applied rules consume is removed from the neuron. In this way, the proposed strategy avoids checking whether all the applied rules consume the same number of spikes from the given neuron. The computation power of RSSN P systems working in the proposed strategy is investigated, and it is proved that such systems characterize the semilinear sets of natural numbers, i.e., such systems are not universal. Furthermore, RSSN P systems with weighted synapses working in the proposed strategy are proved to be Turing universal. These results show that the weight on synapses is a powerful ingredient of RSSN P systems in terms of the computation power, which makes RSSN P systems working in sum spikes consumption strategy become universal from non-universality.

Spiking neural (SN, for short) P systems are a class of computation models inspired from the way in which neurons communicate by exchanging spikes. SN P systems with homogenous neurons and synapses are a new variant of SN P systems, where the spiking and forgetting rules are placed on synapses instead of in neurons and each synapse has the same set of spiking and forgetting rules. Recent studies illustrated that this variant of SN P systems is Turing universal as both number generating and accepting devices. In this note, we prove that SN P systems with homogenous neurons and synapses without the feature of delay are also Turing universal. This result gives a positive answer to an open problem formulated in [K. Jiang, et al. Neurocomputing 171(2016) 1548-1555] “whether SN P systems with homogenous neurons and synapses are Turing universal when the feature of delay is not used”.

Spiking neural (SN, for short) P systems are a class of distributed parallel computing models inspired by the way in which neurons communicate with each other by means of electrical impulses. Recently, a new variant of SN P systems, called SN P systems with homogenous neurons and synapses (HRSSN P systems for short) was proposed, where the spiking and forgetting rules are placed on synapses instead of in neurons and each synapse has the same set of spiking and forgetting rules. This variant of SN P systems has already been proved to be Turing universal as both number generating and accepting devices. In this work, we consider the problem of looking for small universal HRSSN P systems. Specifically, a universal HRRSN P system with standard rules and weight at most 5 having 70 neurons is constructed as a device of computing functions; as a number generator, we find a universal system with standard rules and weight at most 5 having 71 neurons.

A control strategy on the computations in a one-catalyst P system is provided: the rules are assumed “colored” and in each step only rules of the same “color” are used. Such control leads to Turing universality for one-catalyst P systems with one membrane. Turing universality is also reached for purely catalytic P systems with two catalysts, and for purely catalytic P systems with only one catalyst and cooperating rules working in the so-called terminal mode.

Numerical P systems are a class of P systems inspired both from the structure of living cells and from economics. In this work, we further investigate the generative capacity of numerical P systems as language generators. The families of languages generated by non-enzymatic, by enzymatic, and by purely enzymatic (all programs are enzymatic) numerical P systems working in the sequential mode are compared with the language families in the Chomsky hierarchy. Especially, a characterization of recursively enumerable languages is obtained by using purely enzymatic numerical P systems working in the sequential mode.