An official website of the United States government
Abstract Substantial progresses in protein structure prediction have been made by utilizing deep‐learning and residue‐residue distance prediction since CASP13. Inspired by the advances, we improve our CASP14 MULTICOM protein structure prediction system by incorporating three new components: (a) a new deep learning‐based protein inter‐residue distance predictor to improve template‐free (ab initio) tertiary structure prediction, (b) an enhanced template‐based tertiary structure prediction method, and (c) distance‐based model quality assessment methods empowered by deep learning. In the 2020 CASP14 experiment, MULTICOM predictor was ranked seventh out of 146 predictors in tertiary structure prediction and ranked third out of 136 predictors in inter‐domain structure prediction. The results demonstrate that the template‐free modeling based on deep learning and residue‐residue distance prediction can predict the correct topology for almost all template‐based modeling targets and a majority of hard targets (template‐free targets or targets whose templates cannot be recognized), which is a significant improvement over the CASP13 MULTICOM predictor. Moreover, the template‐free modeling performs better than the template‐based modeling on not only hard targets but also the targets that have homologous templates. The performance of the template‐free modeling largely depends on the accuracy of distance prediction closely related to the quality of multiple sequence alignments. The structural model quality assessment works well on targets for which enough good models can be predicted, but it may perform poorly when only a few good models are predicted for a hard target and the distribution of model quality scores is highly skewed. MULTICOM is available athttps://github.com/jianlin-cheng/MULTICOM_Human_CASP14/tree/CASP14_DeepRank3andhttps://github.com/multicom-toolbox/multicom/tree/multicom_v2.0.
more »
« less
This content will become publicly available on September 1, 2026
AlphaFold model quality self‐assessment improvement via deep graph learning
Abstract In recent years, significant advancements have been made in deep learning‐based computational modeling of proteins, with DeepMind's AlphaFold2 standing out as a landmark achievement. These computationally modeled protein structures not only provide atomic coordinates but also include self‐confidence metrics to assess the relative quality of the modeling, either for individual residues or the entire protein. However, these self‐confidence scores are not always reliable; for instance, poorly modeled regions of a protein may sometimes be assigned high confidence. To address this limitation, we introduce Equivariant Quality Assessment Folding (EQAFold), an enhanced framework that refines the Local Distance Difference Test prediction head of AlphaFold to generate more accurate self‐confidence scores. Our results demonstrate that EQAFold outperforms the standard AlphaFold architecture and recent model quality assessment protocols in providing more reliable confidence metrics. Source code for EQAFold is available athttps://github.com/kiharalab/EQAFold_public.
more »
« less
- Award ID(s):
- 2151678
- PAR ID:
- 10649432
- Publisher / Repository:
- Wiley
- Date Published:
- Journal Name:
- Protein Science
- Volume:
- 34
- Issue:
- 9
- ISSN:
- 0961-8368
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Protein side-chain packing (PSCP), the problem of predicting side-chain conformations given a fixed backbone structure, has important implications in the modeling of structures and interactions. However, despite the groundbreaking progress in protein structure prediction pioneered by AlphaFold, the existing PSCP methods still rely on experimental inputs, and do not leverage AlphaFold-predicted backbone coordinates to enable PSCP at scale. Here, we perform a large-scale benchmarking of the predictive performance of various PSCP methods on public datasets from multiple rounds of the Critical Assessment of Structure Prediction challenges using a diverse set of evaluation metrics. Empirical results demonstrate that the PSCP methods perform well in packing the side-chains with experimental inputs, but they fail to generalize in repacking AlphaFold-generated structures. We additionally explore the effectiveness of leveraging the self-assessment confidence scores from AlphaFold by implementing a backbone confidence-aware integrative approach. While such a protocol often leads to performance improvement by attaining modest yet statistically significant accuracy gains over the AlphaFold baseline, it does not yield consistent and pronounced improvements. Our study highlights the recent advances and remaining challenges in PSCP in the post-AlphaFold era.more » « less
-
Abstract The Membranome database provides comprehensive structural information on single‐pass (i.e., bitopic) membrane proteins from six evolutionarily distant organisms, including protein–protein interactions, complexes, mutations, experimental structures, and models of transmembrane α‐helical dimers. We present a new version of this database, Membranome 3.0, which was significantly updated by revising the set of 5,758 bitopic proteins and incorporating models generated by AlphaFold 2 in the database. The AlphaFold models were parsed into structural domains located at the different membrane sides, modified to exclude low‐confidence unstructured terminal regions and signal sequences, validated through comparison with available experimental structures, and positioned with respect to membrane boundaries. Membranome 3.0 was re‐developed to facilitate visualization and comparative analysis of multiple 3D structures of proteins that belong to a specified family, complex, biological pathway, or membrane type. New tools for advanced search and analysis of proteins, their interactions, complexes, and mutations were included. The database is freely accessible athttps://membranome.org.more » « less
-
Abstract Macromolecular protein complexes carry out most functions in the cell including essential functions required for cell survival. Unfortunately, we lack the subunit composition for all human protein complexes. To address this gap we integrated >25,000 mass spectrometry experiments using a machine learning approach to identify > 15,000 human protein complexes. We show our map of protein complexes is highly accurate and more comprehensive than previous maps, placing ∼75% of human proteins into their physical contexts. We globally characterize our complexes using protein co-variation data (ProteomeHD.2) and identify co-varying complexes suggesting common functional associations. Our map also generates testable functional hypotheses for 472 uncharacterized proteins which we support using AlphaFold modeling. Additionally, we use AlphaFold modeling to identify 511 mutually exclusive protein pairs in hu.MAP3.0 complexes suggesting complexes serve different functional roles depending on their subunit composition. We identify expression as the primary way cells and organisms relieve the conflict of mutually exclusive subunits. Finally, we import our complexes to EMBL-EBI’s Complex Portal (https://www.ebi.ac.uk/complexportal/home) as well as provide complexes through our hu.MAP3.0 web interface (https://humap3.proteincomplexes.org/). We expect our resource to be highly impactful to the broader research community.more » « less
-
Abstract Many proteins are composed of several domains that pack together into a complex tertiary structure. Multidomain proteins can be challenging for protein structure modeling, particularly those for which templates can be found for individual domains but not for the entire sequence. In such cases, homology modeling can generate high quality models of the domains but not for the orientations between domains. Small‐angle X‐ray scattering (SAXS) reports the structural properties of entire proteins and has the potential for guiding homology modeling of multidomain proteins. In this article, we describe a novel multidomain protein assembly modeling method, SAXSDom that integrates experimental knowledge from SAXS with probabilistic Input‐Output Hidden Markov model to assemble the structures of individual domains together. Four SAXS‐based scoring functions were developed and tested, and the method was evaluated on multidomain proteins from two public datasets. Incorporation of SAXS information improved the accuracy of domain assembly for 40 out of 46 critical assessment of protein structure prediction multidomain protein targets and 45 out of 73 multidomain protein targets from the ab initio domain assembly dataset. The results demonstrate that SAXS data can provide useful information to improve the accuracy of domain‐domain assembly. The source code and tool packages are available athttps://github.com/jianlin-cheng/SAXSDom.more » « less
