Robust prediction of protein–protein interactions (PPIs) requires models that generalize beyond the training distribution. Here, we present PLMDA-PPI, a mechanism-aware framework that incorporates biophysical inductive biases into a dual-attention architecture with residue-level contact supervision applied to protein language model embedded geometric representations. This design enables the model to jointly learn global interaction likelihoods and the residue pairs mediating interactions, producing interpretable, mechanism-grounded predictions. PLMDA-PPI demonstrates strong out-of-distribution generalization, substantially outperforming lightweight deep learning models (D-SCRIPT, Topsy-Turvy, TT3D, and TUnA) and achieving comparable or even superior performance to computationally intensive methods (AF2Complex, RF2-Lite, and RF2-PPI) derived from AlphaFold2 and RoseTTAFold2, while requiring orders of magnitude fewer computational resources. These results show that incorporating mechanistic priors as architectural inductive biases enhances generalization, interpretability, and computational efficiency, providing a principled foundation for AI-driven prediction of complex biological interactions.

Accurate prediction of contacting residue pairs between interacting proteins is very useful for structural characterization of protein-protein interactions (PPIs). Although significant improvement has been made in inter-protein contact prediction recently, there is still large room for improving the prediction accuracy. Here we present a new deep learning method referred to as PLMGraph-Inter for inter-protein contact prediction. Specifically, we employ rotationally and translationally invariant geometric graphs obtained from structures of interacting proteins to integrate multiple protein language models, which are successively transformed by graph encoders formed by geometric vector perceptrons and residual networks formed by dimensional hybrid residual blocks to predict inter-protein contacts. Extensive evaluation on multiple test sets illustrates that PLMGraph-Inter outperforms five top inter-protein contact prediction methods, including DeepHomo, GLINTER, CDPred, DeepHomo2 and DRN-1D2D_Inter by large margins. In addition, we also show that the prediction of PLMGraph-Inter can complement the result of AlphaFold-Multimer. Finally, we show leveraging the contacts predicted by PLMGraph-Inter as constraints for protein-protein docking can dramatically improve its performance for protein complex structure prediction.

Understanding how genetic variants impact molecular phenotypes is a key goal of functional genomics, currently hindered by reliance on a single haploid reference genome. Here, we present the EN-TEx resource of 1,635 open-access datasets from four donors (∼30 tissues × ∼15 assays). The datasets are mapped to matched, diploid genomes with long-read phasing and structural variants, instantiating a catalog of >1 million allele-specific loci. These loci exhibit coordinated activity along haplotypes and are less conserved than corresponding, non-allele-specific ones. Surprisingly, a deep-learning transformer model can predict the allele-specific activity based only on local nucleotide-sequence context, highlighting the importance of transcription-factor-binding motifs particularly sensitive to variants. Furthermore, combining EN-TEx with existing genome annotations reveals strong associations between allele-specific and GWAS loci. It also enables models for transferring known eQTLs to difficult-to-profile tissues (e.g., from skin to heart). Overall, EN-TEx provides rich data and generalizable models for more accurate personal functional genomics.

The knowledge of contacting residue pairs between interacting proteins is very useful for the structural characterization of protein-protein interactions (PPIs). However, accurately identifying the tens of contacting ones from hundreds of thousands of inter-protein residue pairs is extremely challenging, and performances of the state-of-the-art inter-protein contact prediction methods are still quite limited. In this study, we developed a deep learning method for inter-protein contact prediction, which is referred to as DRN-1D2D_Inter. Specifically, we employed pretrained protein language models to generate structural information-enriched input features to residual networks formed by dimensional hybrid residual blocks to perform inter-protein contact prediction. Extensively bechmarking DRN-1D2D_Inter on multiple datasets, including both heteromeric PPIs and homomeric PPIs, we show DRN-1D2D_Inter consistently and significantly outperformed two state-of-the-art inter-protein contact prediction methods, including GLINTER and DeepHomo, although both the latter two methods leveraged the native structures of interacting proteins in the prediction, and DRN-1D2D_Inter made the prediction purely from sequences. We further show that applying the predicted contacts as constraints for protein-protein docking can significantly improve its performance for protein complex structure prediction.

AlphaFold2 can predict protein complex structures as long as a multiple sequence alignment (MSA) of the interologs of the target protein-protein interaction (PPI) can be provided. In this study, a simplified phylogeny-based approach was applied to generate the MSA of interologs, which was then used as the input to AlphaFold2 for protein complex structure prediction. In this extensively benchmarked protocol on nonredundant PPI dataset, including 107 bacterial PPIs and 442 eukaryotic PPIs, we show complex structures of 79.5% of the bacterial PPIs and 49.8% of the eukaryotic PPIs can be successfully predicted, which yielded significantly better performance than the application of MSA of interologs prepared by two existing approaches. Considering PPIs may not be conserved in species with long evolutionary distances, we further restricted interologs in the MSA to different taxonomic ranks of the species of the target PPI in protein complex structure prediction. We found that the success rates can be increased to 87.9% for the bacterial PPIs and 56.3% for the eukaryotic PPIs if interologs in the MSA are restricted to a specific taxonomic rank of the species of each target PPI. Finally, we show that the optimal taxonomic ranks for protein complex structure prediction can be selected with the application of the predicted template modeling (TM) scores of the output models.

Deep residual learning has shown great success in protein contact prediction. In this study, a new deep residual learning-based protein contact prediction model was developed. Comparing with previous models, a new type of residual block hybridizing 1D and 2D convolutions was designed to increase the effective receptive field of the residual network, and a new loss function emphasizing the easily misclassified residue pairs was proposed to enhance the model training. The developed protein contact prediction model referred to as DRN-1D2D was first evaluated on 105 CASP11 targets, 76 CAMEO hard targets and 398 membrane proteins together with two in house-developed reference models based on either the standard 2D residual block or the traditional BCE loss function, from which we confirmed that both the dimensional hybrid residual block and the singularity enhanced loss function can be employed to improve the model performance for protein contact prediction. DRN-1D2D was further evaluated on 39 CASP13 and CASP14 free modeling targets together with the two reference models and six state-of-the-art protein contact prediction models including DeepCov, DeepCon, DeepConPred2, SPOT-Contact, RaptorX-Contact and TripleRes. The result shows that DRN-1D2D consistently achieved the best performance among all these models.

Direct coupling analysis (DCA) has been widely used to infer evolutionary coupled residue pairs from the multiple sequence alignment (MSA) of homologous sequences. However, effectively selecting residue pairs with significant evolutionary couplings according to the result of DCA is a non-trivial task. In this study, we developed a general statistical framework for significant evolutionary coupling detection, referred to as irreproducible discovery rate (IDR)-DCA, which is based on reproducibility analysis of the coupling scores obtained from DCA on manually created MSA replicates. IDR-DCA was applied to select residue pairs for contact prediction for monomeric proteins, protein-protein interactions and monomeric RNAs, in which three different versions of DCA were applied. We demonstrated that with the application of IDR-DCA, the residue pairs selected using a universal threshold always yielded stable performance for contact prediction. Comparing with the application of carefully tuned coupling score cutoffs, IDR-DCA always showed better performance. The robustness of IDR-DCA was also supported through the MSA downsampling analysis. We further demonstrated the effectiveness of applying constraints obtained from residue pairs selected by IDR-DCA to assist RNA secondary structure prediction.

Enhancers are important non-coding elements, but they have traditionally been hard to characterize experimentally. The development of massively parallel assays allows the characterization of large numbers of enhancers for the first time. Here, we developed a framework using Drosophila STARR-seq to create shape-matching filters based on meta-profiles of epigenetic features. We integrated these features with supervised machine-learning algorithms to predict enhancers. We further demonstrated that our model could be transferred to predict enhancers in mammals. We comprehensively validated the predictions using a combination of in vivo and in vitro approaches, involving transgenic assays in mice and transduction-based reporter assays in human cell lines (153 enhancers in total). The results confirmed that our model can accurately predict enhancers in different species without re-parameterization. Finally, we examined the transcription factor binding patterns at predicted enhancers versus promoters. We demonstrated that these patterns enable the construction of a secondary model that effectively distinguishes enhancers and promoters. Supervised machine-learning models trained using Drosophila epigenetic and STARR-seq data can be transferred to predict mouse and human enhancers.

A key issue in drug design is how population variation affects drug efficacy by altering binding affinity (BA) in different individuals, an essential consideration for government regulators. Ideally, we would like to evaluate the BA perturbations of millions of single-nucleotide variants (SNVs). However, only hundreds of protein-drug complexes with SNVs have experimentally characterized BAs, constituting too small a gold standard for straightforward statistical model training. Thus, we take a hybrid approach: using physically based calculations to bootstrap the parameterization of a full model. In particular, we do 3D structure-based docking on ∼10,000 SNVs modifying known protein-drug complexes to construct a pseudo gold standard. Then we use this augmented set of BAs to train a statistical model combining structure, ligand and sequence features and illustrate how it can be applied to millions of SNVs. Finally, we show that our model has good cross-validated performance (97% AUROC) and can also be validated by orthogonal ligand-binding data.

Data science allows the extraction of practical insights from large-scale data. Here, we contextualize it as an umbrella term, encompassing several disparate subdomains. We focus on how genomics fits as a specific application subdomain, in terms of well-known 3V data and 4M process frameworks (volume-velocity-variety and measurement-mining-modeling-manipulation, respectively). We further analyze the technical and cultural exports and imports between genomics and other data-science subdomains (e.g., astronomy). Finally, we discuss how data value, privacy, and ownership are pressing issues for data science applications, in general, and are especially relevant to genomics, due to the persistent nature of DNA.

Protein-peptide interactions play a crucial role in a variety of cellular processes. The protein-peptide complex structure is a key to understand the mechanisms underlying protein-peptide interactions and is critical for peptide therapeutic development. We present a user-friendly protein-peptide docking server, MDockPeP. Starting from a peptide sequence and a protein receptor structure, the MDockPeP Server globally docks the all-atom, flexible peptide to the protein receptor. The produced modes are then evaluated with a statistical potential-based scoring function, ITScorePeP. This method was systematically validated using the peptiDB benchmarking database. At least one near-native peptide binding mode was ranked among top 10 (or top 500) in 59% (85%) of the bound cases, and in 40.6% (71.9%) of the challenging unbound cases. The server can be used for both protein-peptide complex structure prediction and initial-stage sampling of the protein-peptide binding modes for other docking or simulation methods. MDockPeP Server is freely available at . (c) 2018 Wiley Periodicals, Inc.

Despite progress in defining genetic risk for psychiatric disorders, their molecular mechanisms remain elusive. Addressing this, the PsychENCODE Consortium has generated a comprehensive online resource for the adult brain across 1866 individuals. The PsychENCODE resource contains ~79,000 brain-active enhancers, sets of Hi-C linkages, and topologically associating domains; single-cell expression profiles for many cell types; expression quantitative-trait loci (QTLs); and further QTLs associated with chromatin, splicing, and cell-type proportions. Integration shows that varying cell-type proportions largely account for the cross-population variation in expression (with >88% reconstruction accuracy). It also allows building of a gene regulatory network, linking genome-wide association study variants to genes (e.g., 321 for schizophrenia). We embed this network into an interpretable deep-learning model, which improves disease prediction by ~6-fold versus polygenic risk scores and identifies key genes and pathways in psychiatric disorders.

Protein flexibility plays an important role in carrying out protein functions. However, computational modeling of protein flexibility in binding events is of great challenge because proteins have large internal degrees of freedom. In this chapter, the computational methods for modeling protein flexibility in protein-ligand docking and protein-protein docking are classified into three groups, soft docking, induced-fit docking and ensemble docking; each class is reviewed in detail. The chapter has also reviewed the methods for modeling protein flexibility in the post-docking stage of protein-protein docking and protein-ligand docking, and the recent progress for modeling peptide flexibility in the emerging protein-peptide docking.

Summary Genome-wide proximity ligation based assays like Hi-C have opened a window to the 3D organization of the genome. In so doing, they present data structures that are different from conventional 1D signal tracks. To exploit the 2D nature of Hi-C contact maps, matrix techniques like spectral analysis are particularly useful. Here, we present HiC-spector, a collection of matrix-related functions for analyzing Hi-C contact maps. In particular, we introduce a novel reproducibility metric for quantifying the similarity between contact maps based on spectral decomposition. The metric successfully separates contact maps mapped from Hi-C data coming from biological replicates, pseudo-replicates and different cell types. Availability Source code in Julia and the documentation of HiC-spector can be freely obtained at https://github.com/gersteinlab/HiC_spector Contact pi@gersteinlab.org

Peptides mediate up to 40 % of protein-protein interactions in a variety of cellular processes and are also attractive drug candidates. Thus, predicting peptide binding sites on the given protein structure is of great importance for mechanistic investigation of protein-peptide interactions and peptide therapeutics development. In this chapter, we describe the usage of our web server, referred to as ACCLUSTER, for peptide binding site prediction for a given protein structure. ACCLUSTER is freely available for users without registration at http://zougrouptoolkit.missouri.edu/accluster .

The growing number of protein-ligand complex structures, particularly the structures of proteins co-bound with different ligands, in the Protein Data Bank helps us tackle two major challenges in molecular docking studies: the protein flexibility and the scoring function. Here, we introduced a systematic strategy by using the information embedded in the known protein-ligand complex structures to improve both binding mode and binding affinity predictions. Specifically, a ligand similarity calculation method was employed to search a receptor structure with a bound ligand sharing high similarity with the query ligand for the docking use. The strategy was applied to the two datasets (HSP90 and MAP4K4) in recent D3R Grand Challenge 2015. In addition, for the HSP90 dataset, a system-specific scoring function (ITScore2_hsp90) was generated by recalibrating our statistical potential-based scoring function (ITScore2) using the known protein-ligand complex structures and the statistical mechanics-based iterative method. For the HSP90 dataset, better performances were achieved for both binding mode and binding affinity predictions comparing with the original ITScore2 and with ensemble docking. For the MAP4K4 dataset, although there were only eight known protein-ligand complex structures, our docking strategy achieved a comparable performance with ensemble docking. Our method for receptor conformational selection and iterative method for the development of system-specific statistical potential-based scoring functions can be easily applied to other protein targets that have a number of protein-ligand complex structures available to improve predictions on binding.

Protein-peptide interactions play an important role in many cellular processes. In silico prediction of protein-peptide complex structure is highly desirable for mechanistic investigation of these processes and for therapeutic design. However, predicting all-atom structures of protein-peptide complexes without any knowledge about the peptide binding site and the bound peptide conformation remains a big challenge. Here, we present a docking-based method for predicting protein-peptide complex structures, referred to as MDockPeP, which starts with the peptide sequence and globally docks the all-atom, flexible peptide onto the protein structure. MDockPeP was tested on the peptiDB benchmarking database using both bound and unbound protein structures. The results show that MDockPeP successfully generated near-native peptide binding modes in 95.0% of the bound docking cases and in 92.2% of the unbound docking cases. The performance is significantly better than other existing docking methods. MDockPeP is computationally efficient and suitable for large-scale applications.

ABSTRACTProtein‐protein interactions are either through direct contacts between two binding partners or mediated by structural waters. Both direct contacts and water‐mediated interactions are crucial to the formation of a protein‐protein complex. During the recent CAPRI rounds, a novel parallel searching strategy for predicting water‐mediated interactions is introduced into our protein‐protein docking method, MDockPP. Briefly, a FFT‐based docking algorithm is employed in generating putative binding modes, and an iteratively derived statistical potential‐based scoring function, ITScorePP, in conjunction with biological information is used to assess and rank the binding modes. Up to 10 binding modes are selected as the initial protein‐protein complex structures for MD simulations in explicit solvent. Water molecules near the interface are clustered based on the snapshots extracted from independent equilibrated trajectories. Then, protein‐ligand docking is employed for a parallel search for water molecules near the protein‐protein interface. The water molecules generated by ligand docking and the clustered water molecules generated by MD simulations are merged, referred to as the predicted structural water molecules. Here, we report the performance of this protocol for CAPRI rounds 28–29 and 31–35 containing 20 valid docking targets and 11 scoring targets. In the docking experiments, we predicted correct binding modes for nine targets, including one high‐accuracy, two medium‐accuracy, and six acceptable predictions. Regarding the two targets for the prediction of water‐mediated interactions, we achieved models ranked as “excellent” in accordance with the CAPRI evaluation criteria; one of these two targets is considered as a difficult target for structural water prediction. Proteins 2017; 85:424–434. © 2016 Wiley Periodicals, Inc.

ABSTRACTWe present the results for CAPRI Round 30, the first joint CASP‐CAPRI experiment, which brought together experts from the protein structure prediction and protein–protein docking communities. The Round comprised 25 targets from amongst those submitted for the CASP11 prediction experiment of 2014. The targets included mostly homodimers, a few homotetramers, and two heterodimers, and comprised protein chains that could readily be modeled using templates from the Protein Data Bank. On average 24 CAPRI groups and 7 CASP groups submitted docking predictions for each target, and 12 CAPRI groups per target participated in the CAPRI scoring experiment. In total more than 9500 models were assessed against the 3D structures of the corresponding target complexes. Results show that the prediction of homodimer assemblies by homology modeling techniques and docking calculations is quite successful for targets featuring large enough subunit interfaces to represent stable associations. Targets with ambiguous or inaccurate oligomeric state assignments, often featuring crystal contact‐sized interfaces, represented a confounding factor. For those, a much poorer prediction performance was achieved, while nonetheless often providing helpful clues on the correct oligomeric state of the protein. The prediction performance was very poor for genuine tetrameric targets, where the inaccuracy of the homology‐built subunit models and the smaller pair‐wise interfaces severely limited the ability to derive the correct assembly mode. Our analysis also shows that docking procedures tend to perform better than standard homology modeling techniques and that highly accurate models of the protein components are not always required to identify their association modes with acceptable accuracy. Proteins 2016; 84(Suppl 1):323–348. © 2016 The Authors Proteins: Structure, Function, and Bioinformatics Published by Wiley Periodicals, Inc.

Molecular docking refers to computational methods for the prediction of the binding mode and binding affinity between two molecules. Over decades of development, protein–ligand docking methods have been widely used for in silico screening of molecular libraries for drug candidates, serving as a valuable tool in structure-based drug design. MDock is a protein–ligand docking suite originally released from our laboratory in 2007, which incorporates the iteratively derived knowledge-based scoring function and the ensemble docking method. In this chapter, we describe the methodology and usage of MDock for molecular docking and in silico screening. The MDock suite is freely available to academic users through applications at http://zoulab.dalton.missouri.edu/mdock.htm .