Classification of domains in predicted structures of the human proteome

Abstract

Recent advances in protein structure prediction have generated accurate structures of previously uncharacterized human proteins. Identifying domains in these predicted structures and classifying them into an evolutionary hierarchy can reveal biological insights. Here, we describe the detection and classification of domains from the human proteome. Our classification indicates that only 62% of residues are located in globular domains. We further classify these globular domains and observe that the majority (65%) can be classified among known folds by sequence, with a smaller fraction (33%) requiring structural data to refine the domain boundaries and/or to support their homology. A relatively small number (966 domains) cannot be confidently assigned using our automatic pipelines, thus demanding manual inspection. We classify 47,576 domains, of which only 23% have been included in experimental structures. A portion (6.3%) of these classified globular domains lack sequence-based annotation in InterPro. A quarter (23%) have not been structurally modeled by homology, and they contain 2,540 known disease-causing single amino acid variations whose pathogenesis can now be inferred using AF models. A comparison of classified domains from a series of model organisms revealed expansions of several immune response-related domains in humans and a depletion of olfactory receptors. Finally, we use this classification to expand well-known protein families of biological significance. These classifications are presented on the ECOD website ( http://prodata.swmed.edu/ecod/index_human.php ).

ECOD domain classification of 48 whole proteomes from AlphaFold Structure Database using DPAM

“…Because of this difference, we more closely examined the wellassigned domains from those proteomes and enumerated the commonalities and differences among the and are more enriched in mammals and zebrafish. We found a similar enrichment of zinc-binding domains in our earlier classification of the predicted structures of the human proteome [27]. The EGF-like domain (Fig 3G), a small disulfide-bonded pair of β-sheets, is commonly an extracellular signaling factor but is implicated in a variety of roles in the extracellular environment, which likely explains their relative lack in bacteria and fungi.…”

Protein-Folding Chaperones Predict Structure-Function Relationships and Cancer Risk inBRCA1Mutation Carriers

“…A limitation of this study is that the target variation set must map to globular domains to allow interrogation of changes in chaperone binding across different variants. >60% of residues in the human proteome are found in globular domains 12 , thus our approach has broad applicability for evaluating structure-function and genotype-phenotype relationships in humans. One can also envision instances where decreased chaperone binding to variants is informative, such as by binding of a co-factor or compound that induces protein stabilization, or a mutation that switches an intrinsically disordered protein to a well-structured protein.…”

“…Protein folding is a fundamental process frequently disrupted by pathogenic mutations. Most proteins fold into specific three-dimensional structures encoded by the primary amino acid sequence [9][10][11][12] . Globular domains are the most well-characterized fundamental units of protein structure, which generally fold independently from the rest of the protein.…”

DrugDomain: the evolutionary context of drugs and small molecules bound to domains

“…AlphaFold (AF), a recently developed deep learning method, demonstrated the capability to predict protein structure with atomic-level accuracy and has thus become an indispensable tool in structural biology (Jumper et al 2021). Utilizing AF models, ECOD became one of the first databases to incorporate domain classification both for the entire human proteome (Schaeffer et al 2023) and the whole proteomes of 48 model organisms (Schaeffer et al 2024). AlphaFold has significantly expanded the available tools for computational structural biology related to drug discovery, target prediction, protein-protein and protein-ligand interaction, and prediction of complex structures (Akdel et al 2022; Medvedev et al 2023a; Yang et al 2023).…”

“…In these cases, we applied the AlphaFill algorithm (Hekkelman et al 2023) that uses sequence and structure similarity to retrieve small molecules and ions from experimentally determined structures to predicted protein models by AlphaFold. Using AlphaFill models we identify residues whose atoms are located within 5Å of the DrugBank molecule’s atoms of interest (if present) and map these residues to ECOD domains identified for the whole human proteome using AlphaFold models (Schaeffer et al 2023; Schaeffer et al 2024). If the molecule of interest is not present in the AlphaFold model, all other present ligands are considered as such.…”