Learning the local landscape of protein structures with convolutional neural networks

Kulikova, Anastasiya V.; Diaz, Daniel J.; Loy, James M.; Ellington, Andrew D.; Wilke, Claus O.

doi:10.1007/s10867-021-09593-6

Cited by 15 publications

(37 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such sequences are useful as references for bioinformatic processing in, for instance, alignments and contig assembly, for detection of hypermutants, gene detection and annotation, and for representing simplified views and data from complex populations (Rose and Korber, 2000;Lee, 2003;Seah et al, 2020;Domingo et al, 2021;Frith et al, 2021;Kulikova et al, 2021;Zhang et al, 2021). Consensus sequences have also been used in studies of protein functions, binding, and vaccine designs (Novitsky et al, 2002;Gao et al, 2005;Nickle et al, 2007;Yan et al, 2007;Sternke et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

Updated HIV-1 Consensus Sequences Change but Stay Within Similar Distance From Worldwide Samples

2022

View full text Add to dashboard Cite

HIV consensus sequences are used in various bioinformatic, evolutionary, and vaccine related research. Since the previous HIV-1 subtype and CRF consensus sequences were constructed in 2002, the number of publicly available HIV-1 sequences have grown exponentially, especially from non-EU and US countries. Here, we reconstruct 90 new HIV-1 subtype and CRF consensus sequences from 3,470 high-quality, representative, full genome sequences in the LANL HIV database. While subtypes and CRFs are unevenly spread across the world, in total 89 countries were represented. For consensus sequences that were based on at least 20 genomes, we found that on average 2.3% (range 0.8–10%) of the consensus genome site states changed from 2002 to 2021, of which about half were nucleotide state differences and the rest insertions and deletions. Interestingly, the 2021 consensus sequences were shorter than in 2002, and compared to 4,674 HIV-1 worldwide genome sequences, the 2021 consensuses were somewhat closer to the worldwide genome sequences, i.e., showing on average fewer nucleotide state differences. Some subtypes/CRFs have had limited geographical spread, and thus sampling of subtypes/CRFs is uneven, at least in part, due to the epidemiological dynamics. Thus, taken as a whole, the 2021 consensus sequences likely are good representations of the typical subtype/CRF genome nucleotide states. The new consensus sequences are available at the LANL HIV database.

show abstract

Section: Introductionmentioning

confidence: 99%

Updated HIV-1 Consensus Sequences Change but Stay Within Similar Distance From Worldwide Samples

2022

View full text Add to dashboard Cite

show abstract

“…Model confidence is defined as the probability of the top-1 prediction at a site, and it correlates well with model accuracy (Supplementary Fig. S4 and prior work 17 ). Therefore, we used it here as an approximation of model accuracy, which is not defined for individual sites (a site either is or is not predicted correctly).…”

Section: Resultsmentioning

confidence: 98%

“…We first assessed each model separately, using a test dataset we had previously used in studying the 3D CNN model 17 . Our test set is derived from the PSICOV dataset 29 , which consists of 150 well studied protein structures commonly used for covariation analyses.…”

Section: Testing Individual Modelsmentioning

confidence: 99%

Two sequence- and two structure-based ML models have learned different aspects of protein biochemistry

Kulikova

Diaz

Chen

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Deep learning models are seeing increased use as methods to predict mutational effects or allowed mutations at various sites in proteins. The models commonly used for these purposes include large language models (LLMs) and 3D Convolutional Neural Networks (CNNs). These two model types have very different architectures and are trained on different representations of proteins. LLMs make use of the transformer architecture and are trained purely on protein sequences whereas 3D CNNs are trained on voxelized representations of local protein structure. While comparable overall prediction accuracies have been reported for both types of models, it is not known to what extent these models make comparable specific predictions and/or generalize protein biochemistry in similar ways. Here, we perform a systematic comparison of two LLMs and one 3D CNN model and show that the different model types have distinct strengths and weaknesses. The overall prediction accuracies are largely uncorrelated between sequence and structure based models. Overall, the 3D CNN model is better at predicting buried aliphatic and hydrophobic residues whereas the LLMs are better at predicting solvent-exposed polar and charged amino acids. A combined model that takes the individual model predictions as input can leverage these individual model strengths and results in significantly improved overall prediction accuracy.

show abstract

“…Other attempts tried to encode more information in the input to CNNs, e.g. CNN protein landscape [64] , which amongst others encoded side chain atoms, partial charges and solvent accessibility reaching 60 % NSR, and TIMED [65] , which besides included reimplementations of several CNN-based protein design methods.…”

Section: The Deep Learning Era Of Protein Sequence and Structure Gene...mentioning

confidence: 99%

From sequence to function through structure: Deep learning for protein design

Ferruz¹,

Heinzinger²,

Akdel³

et al. 2023

Computational and Structural Biotechnology Journal

View full text Add to dashboard Cite

Learning the local landscape of protein structures with convolutional neural networks

Cited by 15 publications

References 40 publications

Updated HIV-1 Consensus Sequences Change but Stay Within Similar Distance From Worldwide Samples

Updated HIV-1 Consensus Sequences Change but Stay Within Similar Distance From Worldwide Samples

Two sequence- and two structure-based ML models have learned different aspects of protein biochemistry

From sequence to function through structure: Deep learning for protein design

Contact Info

Product

Resources

About