Dataset Information

Predicting accurate ab initio DNA electron densities with equivariant neural networks.

ABSTRACT: One of the fundamental limitations of accurately modeling biomolecules like DNA is the inability to perform quantum chemistry calculations on large molecular structures. We present a machine learning model based on an equivariant Euclidean neural network framework to obtain accurate ab initio electron densities for arbitrary DNA structures that are much too large for conventional quantum methods. The model is trained on representative B-DNA basepair steps that capture both base pairing and base stacking interactions. The model produces accurate electron densities for arbitrary B-DNA structures with typical errors of less than 1%. Crucially, the error does not increase with system size, which suggests that the model can extrapolate to large DNA structures with negligible loss of accuracy. The model also generalizes reasonably to other DNA structural motifs such as the A- and Z-DNA forms, despite being trained on only B-DNA configurations. The model is used to calculate electron densities of several large-scale DNA structures, and we show that the computational scaling for this model is essentially linear. We also show that this machine learning electron density model can be used to calculate accurate electrostatic potentials for DNA. These electrostatic potentials produce more accurate results compared with classical force fields and do not show the usual deficiencies at short range.

SUBMITTER: Lee AJ

PROVIDER: S-EPMC9674991 | biostudies-literature | 2022 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Predicting accurate ab initio DNA electron densities with equivariant neural networks.

Lee Alex J AJ Rackers Joshua A JA Bricker William P WP

Biophysical journal 20220903 20

One of the fundamental limitations of accurately modeling biomolecules like DNA is the inability to perform quantum chemistry calculations on large molecular structures. We present a machine learning model based on an equivariant Euclidean neural network framework to obtain accurate ab initio electron densities for arbitrary DNA structures that are much too large for conventional quantum methods. The model is trained on representative B-DNA basepair steps that capture both base pairing and base ...[more]

PMID: 36057785

Similar Datasets

Project description:BACKGROUND:Proteins, especially larger ones, are often composed of individual evolutionary units, domains, which have their own function and structural fold. Predicting domains is an important intermediate step in protein analyses, including the prediction of protein structures. RESULTS:We describe novel systems for the prediction of protein domain boundaries powered by Recursive Neural Networks. The systems rely on a combination of primary sequence and evolutionary information, predictions of structural features such as secondary structure, solvent accessibility and residue contact maps, and structural templates, both annotated for domains (from the SCOP dataset) and unannotated (from the PDB). We gauge the contribution of contact maps, and PDB and SCOP templates independently and for different ranges of template quality. We find that accurately predicted contact maps are informative for the prediction of domain boundaries, while the same is not true for contact maps predicted ab initio. We also find that gap information from PDB templates is informative, but, not surprisingly, less than SCOP annotations. We test both systems trained on templates of all qualities, and systems trained only on templates of marginal similarity to the query (less than 25% sequence identity). While the first batch of systems produces near perfect predictions in the presence of fair to good templates, the second batch outperforms or match ab initio predictors down to essentially any level of template quality.We test all systems in 5-fold cross-validation on a large non-redundant set of multi-domain and single domain proteins. The final predictors are state-of-the-art, with a template-less prediction boundary recall of 50.8% (precision 38.7%) within +/- 20 residues and a single domain recall of 80.3% (precision 78.1%). The SCOP-based predictors achieve a boundary recall of 74% (precision 77.1%) again within +/- 20 residues, and classify single domain proteins as such in over 85% of cases, when we allow a mix of bad and good quality templates. If we only allow marginal templates (max 25% sequence identity to the query) the scores remain high, with boundary recall and precision of 59% and 66.3%, and 80% of all single domain proteins predicted correctly. CONCLUSION:The systems presented here may prove useful in large-scale annotation of protein domains in proteins of unknown structure. The methods are available as public web servers at the address: http://distill.ucd.ie/shandy/ and we plan on running them on a multi-genomic scale and make the results public in the near future.

Project description:Predictive modeling of toxicity is a crucial step in the drug discovery pipeline. It can help filter out molecules with a high probability of failing in the early stages of de novo drug design. Thus, several machine learning (ML) models have been developed to predict the toxicity of molecules by combining classical ML techniques or deep neural networks with well-known molecular representations such as fingerprints or 2D graphs. But the more natural, accurate representation of molecules is expected to be defined in physical 3D space like in ab initio methods. Recent studies successfully used equivariant graph neural networks (EGNNs) for representation learning based on 3D structures to predict quantum-mechanical properties of molecules. Inspired by this, we investigated the performance of EGNNs to construct reliable ML models for toxicity prediction. We used the equivariant transformer (ET) model in TorchMD-NET for this. Eleven toxicity data sets taken from MoleculeNet, TDCommons, and ToxBenchmark have been considered to evaluate the capability of ET for toxicity prediction. Our results show that ET adequately learns 3D representations of molecules that can successfully correlate with toxicity activity, achieving good accuracies on most data sets comparable to state-of-the-art models. We also test a physicochemical property, namely, the total energy of a molecule, to inform the toxicity prediction with a physical prior. However, our work suggests that these two properties can not be related. We also provide an attention weight analysis for helping to understand the toxicity prediction in 3D space and thus increase the explainability of the ML model. In summary, our findings offer promising insights considering 3D geometry information via EGNNs and provide a straightforward way to integrate molecular conformers into ML-based pipelines for predicting and investigating toxicity prediction in physical space. We expect that in the future, especially for larger, more diverse data sets, EGNNs will be an essential tool in this domain.

Project description:Ab initio gene identification in the genomic sequence of Drosophila melanogaster was obtained using (human gene predictor) and Fgenesh programs that have organism-specific parameters for human, Drosophila, plants, yeast, and nematode. We did not use information about cDNA/EST in most predictions to model a real situation for finding new genes because information about complete cDNA is often absent or based on very small partial fragments. We investigated the accuracy of gene prediction on different levels and designed several schemes to predict an unambiguous set of genes (annotation CGG1), a set of reliable exons (annotation CGG2), and the most complete set of exons (annotation CGG3). For 49 genes, protein products of which have clear homologs in protein databases, predictions were recomputed by Fgenesh+ program. The first annotation serves as the optimal computational description of new sequence to be presented in a database. Reliable exons from the second annotation serve as good candidates for selecting the PCR primers for experimental work for gene structure verification. Our results shows that we can identify approximately 90% of coding nucleotides with 20% false positives. At the exon level we accurately predicted 65% of exons and 89% including overlapping exons with 49% false positives. Optimizing accuracy of prediction, we designed a gene identification scheme using Fgenesh, which provided sensitivity (Sn) = 98% and specificity (Sp) = 86% at the base level, Sn = 81% (97% including overlapping exons) and Sp = 58% at the exon level and Sn = 72% and Sp = 39% at the gene level (estimating sensitivity on std1 set and specificity on std3 set). In general, these results showed that computational gene prediction can be a reliable tool for annotating new genomic sequences, giving accurate information on 90% of coding sequences with 14% false positives. However, exact gene prediction (especially at the gene level) needs additional improvement using gene prediction algorithms. The program was also tested for predicting genes of human Chromosome 22 (the last variant of Fgenesh can analyze the whole chromosome sequence). This analysis has demonstrated that the 88% of manually annotated exons in Chromosome 22 were among the ab initio predicted exons. The suite of gene identification programs is available through the WWW server of Computational Genomics Group at http://genomic.sanger.ac.uk/gf. html.

Dataset Information

Predicting accurate ab initio DNA electron densities with equivariant neural networks.

Publications

Predicting accurate ab initio DNA electron densities with equivariant neural networks.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets