Foundation models in molecular biology.

Determining correlations between molecules at various levels is an important topic in molecular biology. Large language models have demonstrated a remarkable ability to capture correlations from large amounts of data in the field of natural language processing as well as image generation, and correlations captured from data using large language models can also be applicable to solving a wide range of specific tasks, hence large language models are also referred to as foundation models. The massive amount of data that exists in the field of molecular biology provides an excellent basis for the development of foundation models, and the recent emergence of foundation models in the field of molecular biology has really pushed the entire field forward. We summarize the foundation models developed based on RNA sequence data, DNA sequence data, protein sequence data, single-cell transcriptome data, and spatial transcriptome data respectively, and further discuss the research directions for the development of foundation models in molecular biology.

Protein language model-embedded geometric graphs power inter-protein contact prediction.

Accurate prediction of contacting residue pairs between interacting proteins is very useful for structural characterization of protein-protein interactions. Although significant improvement has been made in inter-protein contact prediction recently, there is still a 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.

Improved inter-protein contact prediction using dimensional hybrid residual networks and protein language models.

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.

Protein complex structure prediction powered by multiple sequence alignments of interologs from multiple taxonomic ranks and AlphaFold2.

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.

A reproducibility analysis-based statistical framework for residue-residue evolutionary coupling detection.

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.

Improved protein contact prediction using dimensional hybrid residual networks and singularity enhanced loss function.

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.