Petascale Homology Search for Structure Prediction

NEAR: Neural Embeddings for Amino acid Relationships

Protein language models (PLMs) have recently demonstrated potential to supplant classical protein database search methods based on sequence alignment, but are slower than common alignment-based tools and appear to be prone to a high rate of false labeling. Here, we present NEAR, a method based on neural representation learning that is designed to improve both speed and accuracy of search for likely homologs in a large protein sequence database. NEAR's ResNet embedding model is trained using contrastive learning guided by trusted sequence alignments. It computes per-residue embeddings for target and query protein sequences, and identifies alignment candidates with a pipeline consisting of residue-level k-NN search and a simple neighbor aggregation scheme. Tests on a benchmark consisting of trusted remote homologs and randomly shuffled decoy sequences reveal that NEAR substantially improves accuracy relative to state-of-the-art PLMs, with lower memory requirements and faster embedding and search speed. While these results suggest that the NEAR model may be useful for standalone homology detection with increased sensitivity over standard alignment-based methods, in this manuscript we focus on a more straightforward analysis of the model's value as a high-speed pre-filter for sensitive annotation. In that context, NEAR is at least 5x faster than the pre-filter currently used in the widely-used profile hidden Markov model (pHMM) search tool HMMER3 , and also outperforms the pre-filter used in our fast pHMM tool, nail .

Mapping glycoprotein structure reveals Flaviviridae evolutionary history

Viral glycoproteins drive membrane fusion in enveloped viruses and determine host range, tissue tropism and pathogenesis1. Despite their importance, there is a fragmentary understanding of glycoproteins within the Flaviviridae2, a large virus family that include pathogens such as hepatitis C, dengue and Zika viruses, and numerous other human, animal and emergent viruses. For many flaviviruses the glycoproteins have not yet been identified, for others, such as the hepaciviruses, the molecular mechanisms of membrane fusion remain uncharacterized3. Here we combine phylogenetic analyses with protein structure prediction to survey glycoproteins across the entire Flaviviridae. We find class II fusion systems, homologous to the Orthoflavivirus E glycoprotein in most species, including highly divergent jingmenviruses and large genome flaviviruses. However, the E1E2 glycoproteins of the hepaciviruses, pegiviruses and pestiviruses are structurally distinct, may represent a novel class of fusion mechanism, and are strictly associated with infection of vertebrate hosts. By mapping glycoprotein distribution onto the underlying phylogeny, we reveal a complex evolutionary history marked by the capture of bacterial genes and potentially inter-genus recombination. These insights, made possible through protein structure prediction, refine our understanding of viral fusion mechanisms and reveal the events that have shaped the diverse virology and ecology of the Flaviviridae.

Structure prediction of alternative protein conformations

Proteins are dynamic molecules whose movements result in different conformations with different functions. Neural networks such as AlphaFold2 can predict the structure of single-chain proteins with conformations most likely to exist in the PDB. However, almost all protein structures with multiple conformations represented in the PDB have been used while training these models. Therefore, it is unclear whether alternative protein conformations can be genuinely predicted using these networks, or if they are simply reproduced from memory. Here, we train a structure prediction network, Cfold, on a conformational split of the PDB to generate alternative conformations. Cfold enables efficient exploration of the conformational landscape of monomeric protein structures. Over 50% of experimentally known nonredundant alternative protein conformations evaluated here are predicted with high accuracy (TM-score > 0.8).

Deep learning insights into the architecture of the mammalian egg-sperm fusion synapse

A crucial event in sexual reproduction is when haploid sperm and egg fuse to form a new diploid organism at fertilization. In mammals, direct interaction between egg JUNO and sperm IZUMO1 mediates gamete membrane adhesion, yet their role in fusion remains enigmatic. We used AlphaFold to predict the structure of other extracellular proteins essential for fertilization to determine if they could form a complex that may mediate fusion. We first identified TMEM81, whose gene is expressed by mouse and human spermatids, as a protein having structural homologies with both IZUMO1 and another sperm molecule essential for gamete fusion, SPACA6. Using a set of proteins known to be important for fertilization and TMEM81, we then systematically searched for predicted binary interactions using an unguided approach and identified a pentameric complex involving sperm IZUMO1, SPACA6, TMEM81 and egg JUNO, CD9. This complex is structurally consistent with both the expected topology on opposing gamete membranes and the location of predicted N-glycans not modeled by AlphaFold-Multimer, suggesting that its components could organize into a synapse-like assembly at the point of fusion. Finally, the structural modeling approach described here could be more generally useful to gain insights into transient protein complexes difficult to detect experimentally.

High-throughput prediction of protein conformational distributions with subsampled AlphaFold2

This paper presents an innovative approach for predicting the relative populations of protein conformations using AlphaFold 2, an AI-powered method that has revolutionized biology by enabling the accurate prediction of protein structures. While AlphaFold 2 has shown exceptional accuracy and speed, it is designed to predict proteins' ground state conformations and is limited in its ability to predict conformational landscapes. Here, we demonstrate how AlphaFold 2 can directly predict the relative populations of different protein conformations by subsampling multiple sequence alignments. We tested our method against nuclear magnetic resonance experiments on two proteins with drastically different amounts of available sequence data, Abl1 kinase and the granulocyte-macrophage colony-stimulating factor, and predicted changes in their relative state populations with more than 80% accuracy. Our subsampling approach worked best when used to qualitatively predict the effects of mutations or evolution on the conformational landscape and well-populated states of proteins. It thus offers a fast and cost-effective way to predict the relative populations of protein conformations at even single-point mutation resolution, making it a useful tool for pharmacology, analysis of experimental results, and predicting evolution.