Positional SHAP (PoSHAP) for Interpretation of machine learning models trained from biological sequences.
Ontology highlight
ABSTRACT:
SUBMITTER: Dickinson Q
PROVIDER: S-EPMC8797255 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:BackgroundDecoding human genomic sequences requires comprehensive analysis of DNA sequence functionality. Through computational and experimental approaches, researchers have studied the genotype-phenotype relationship and generate important datasets that help unravel complicated genetic blueprints. Thus, the recently developed artificial intelligence methods can be used to interpret the functions of those DNA sequences.MethodsThis study explores the use of deep learning, particularly pre-trained genomic models like DNA_bert_6 and human_gpt2-v1, in interpreting and representing human genome sequences. Initially, we meticulously constructed multiple datasets linking genotypes and phenotypes to fine-tune those models for precise DNA sequence classification. Additionally, we evaluate the influence of sequence length on classification results and analyze the impact of feature extraction in the hidden layers of our model using the HERV dataset. To enhance our understanding of phenotype-specific patterns recognized by the model, we perform enrichment, pathogenicity and conservation analyzes of specific motifs in the human endogenous retrovirus (HERV) sequence with high average local representation weight (ALRW) scores.ResultsWe have constructed multiple genotype-phenotype datasets displaying commendable classification performance in comparison with random genomic sequences, particularly in the HERV dataset, which achieved binary and multi-classification accuracies and F1 values exceeding 0.935 and 0.888, respectively. Notably, the fine-tuning of the HERV dataset not only improved our ability to identify and distinguish diverse information types within DNA sequences but also successfully identified specific motifs associated with neurological disorders and cancers in regions with high ALRW scores. Subsequent analysis of these motifs shed light on the adaptive responses of species to environmental pressures and their co-evolution with pathogens.ConclusionsThese findings highlight the potential of pre-trained genomic models in learning DNA sequence representations, particularly when utilizing the HERV dataset, and provide valuable insights for future research endeavors. This study represents an innovative strategy that combines pre-trained genomic model representations with classical methods for analyzing the functionality of genome sequences, thereby promoting cross-fertilization between genomics and artificial intelligence.
| S-EPMC11318145 | biostudies-literature
Project description:BackgroundBoltzmann machines are energy-based models that have been shown to provide an accurate statistical description of domains of evolutionary-related protein and RNA families. They are parametrized in terms of local biases accounting for residue conservation, and pairwise terms to model epistatic coevolution between residues. From the model parameters, it is possible to extract an accurate prediction of the three-dimensional contact map of the target domain. More recently, the accuracy of these models has been also assessed in terms of their ability in predicting mutational effects and generating in silico functional sequences.ResultsOur adaptive implementation of Boltzmann machine learning, adabmDCA, can be generally applied to both protein and RNA families and accomplishes several learning set-ups, depending on the complexity of the input data and on the user requirements. The code is fully available at https://github.com/anna-pa-m/adabmDCA . As an example, we have performed the learning of three Boltzmann machines modeling the Kunitz and Beta-lactamase2 protein domains and TPP-riboswitch RNA domain.ConclusionsThe models learned by adabmDCA are comparable to those obtained by state-of-the-art techniques for this task, in terms of the quality of the inferred contact map as well as of the synthetically generated sequences. In addition, the code implements both equilibrium and out-of-equilibrium learning, which allows for an accurate and lossless training when the equilibrium one is prohibitive in terms of computational time, and allows for pruning irrelevant parameters using an information-based criterion.
| S-EPMC8555268 | biostudies-literature
Project description:Thioamide substitutions of the peptide backbone have been shown to stabilize therapeutic and imaging peptides toward proteolysis. In order to rationally design thioamide modifications, we have developed a novel Rosetta custom score function to classify thioamide positional effects on proteolysis in substrates of serine and cysteine proteases. Peptides of interest were docked into proteases using the FlexPepDock application in Rosetta. Docked complexes were modified to contain thioamides parametrized through the creation of custom atom types in Rosetta based on ab intio simulations. Thioamide complexes were simulated, and the resultant structural complexes provided features for machine learning classification as the decomposed values of the Rosetta score function. An ensemble, majority voting model was developed to be a robust predictor of previously unpublished thioamide proteolysis holdout data. Theoretical control simulations with pseudo-atoms that modulate only one physical characteristic of the thioamide show differential effects on prediction accuracy by the optimized voting classification model. These pseudo-atom model simulations, as well as statistical analyses of the full thioamide simulations, implicate steric effects on peptide binding as being primarily responsible for thioamide positional effects on proteolytic resistance.
| S-EPMC7759720 | biostudies-literature
Project description:In biomedical applications of machine learning, relevant information often has a rich structure that is not easily encoded as real-valued predictors. Examples of such data include DNA or RNA sequences, gene sets or pathways, gene interaction or coexpression networks, ontologies, and phylogenetic trees. We highlight recent examples of machine learning models that use structure to constrain model architecture or incorporate structured data into model training. For machine learning in biomedicine, where sample size is limited and model interpretability is crucial, incorporating prior knowledge in the form of structured data can be particularly useful. The area of research would benefit from performant open source implementations and independent benchmarking efforts.
| S-EPMC7308204 | biostudies-literature
Project description:Organic synthesis remains a major challenge in drug discovery. Although a plethora of machine learning models have been proposed as solutions in the literature, they suffer from being opaque black-boxes. It is neither clear if the models are making correct predictions because they inferred the salient chemistry, nor is it clear which training data they are relying on to reach a prediction. This opaqueness hinders both model developers and users. In this paper, we quantitatively interpret the Molecular Transformer, the state-of-the-art model for reaction prediction. We develop a framework to attribute predicted reaction outcomes both to specific parts of reactants, and to reactions in the training set. Furthermore, we demonstrate how to retrieve evidence for predicted reaction outcomes, and understand counterintuitive predictions by scrutinising the data. Additionally, we identify Clever Hans predictions where the correct prediction is reached for the wrong reason due to dataset bias. We present a new debiased dataset that provides a more realistic assessment of model performance, which we propose as the new standard benchmark for comparing reaction prediction models.
| S-EPMC7966799 | biostudies-literature
Project description:Natural compounds constitute a rich resource of potential small molecule therapeutics. While experimental access to this resource is limited due to its vast diversity and difficulties in systematic purification, computational assessment of structural similarity with known therapeutic molecules offers a scalable approach. Here, we assessed functional similarity between natural compounds and approved drugs by combining multiple chemical similarity metrics and physicochemical properties using a machine-learning approach. We computed pairwise similarities between 1410 drugs for training classification models and used the drugs shared protein targets as class labels. The best performing models were random forest which gave an average area under the ROC of 0.9, Matthews correlation coefficient of 0.35, and F1 score of 0.33, suggesting that it captured the structure-activity relation well. The models were then used to predict protein targets of circa 11k natural compounds by comparing them with the drugs. This revealed therapeutic potential of several natural compounds, including those with support from previously published sources as well as those hitherto unexplored. We experimentally validated one of the predicted pair's activities, viz., Cox-1 inhibition by 5-methoxysalicylic acid, a molecule commonly found in tea, herbs and spices. In contrast, another natural compound, 4-isopropylbenzoic acid, with the highest similarity score when considering most weighted similarity metric but not picked by our models, did not inhibit Cox-1. Our results demonstrate the utility of a machine-learning approach combining multiple chemical features for uncovering protein binding potential of natural compounds.
| S-EPMC9071136 | biostudies-literature
Project description:Early detection of breast cancer and its correct stage determination are important for prognosis and rendering appropriate personalized clinical treatment to breast cancer patients. However, despite considerable efforts and progress, there is a need to identify the specific genomic factors responsible for, or accompanying Invasive Ductal Carcinoma (IDC) progression stages, which can aid the determination of the correct cancer stages. We have developed two-class machine-learning classification models to differentiate the early and late stages of IDC. The prediction models are trained with RNA-seq gene expression profiles representing different IDC stages of 610 patients, obtained from The Cancer Genome Atlas (TCGA). Different supervised learning algorithms were trained and evaluated with an enriched model learning, facilitated by different feature selection methods. We also developed a machine-learning classifier trained on the same datasets with training sets reduced data corresponding to IDC driver genes. Based on these two classifiers, we have developed a web-server Duct-BRCA-CSP to predict early stage from late stages of IDC based on input RNA-seq gene expression profiles. The analysis conducted by us also enables deeper insights into the stage-dependent molecular events accompanying IDC progression. The server is publicly available at http://bioinfo.icgeb.res.in/duct-BRCA-CSP.
| S-EPMC7057992 | biostudies-literature
Project description:BackgroundA new algorithm has been developed to enable the interpretation of black box models. The developed algorithm is agnostic to learning algorithm and open to all structural based descriptors such as fragments, keys and hashed fingerprints. The algorithm has provided meaningful interpretation of Ames mutagenicity predictions from both random forest and support vector machine models built on a variety of structural fingerprints.A fragmentation algorithm is utilised to investigate the model's behaviour on specific substructures present in the query. An output is formulated summarising causes of activation and deactivation. The algorithm is able to identify multiple causes of activation or deactivation in addition to identifying localised deactivations where the prediction for the query is active overall. No loss in performance is seen as there is no change in the prediction; the interpretation is produced directly on the model's behaviour for the specific query.ResultsModels have been built using multiple learning algorithms including support vector machine and random forest. The models were built on public Ames mutagenicity data and a variety of fingerprint descriptors were used. These models produced a good performance in both internal and external validation with accuracies around 82%. The models were used to evaluate the interpretation algorithm. Interpretation was revealed that links closely with understood mechanisms for Ames mutagenicity.ConclusionThis methodology allows for a greater utilisation of the predictions made by black box models and can expedite further study based on the output for a (quantitative) structure activity model. Additionally the algorithm could be utilised for chemical dataset investigation and knowledge extraction/human SAR development.
| S-EPMC3997921 | biostudies-literature
Project description:Background and objectivesAccurate predictions of haemoglobin (Hb) deferral for whole-blood donors could aid blood banks in reducing deferral rates and increasing efficiency and donor motivation. Complex models are needed to make accurate predictions, but predictions must also be explainable. Before the implementation of a prediction model, its impact on the blood supply should be estimated to avoid shortages.Materials and methodsDonation visits between October 2017 and December 2021 were selected from Sanquin's database system. The following variables were available for each visit: donor sex, age, donation start time, month, number of donations in the last 24 months, most recent ferritin level, days since last ferritin measurement, Hb at nth previous visit (n between 1 and 5), days since the nth previous visit. Outcome Hb deferral has two classes: deferred and not deferred. Support vector machines were used as prediction models, and SHapley Additive exPlanations values were used to quantify the contribution of each variable to the model predictions. Performance was assessed using precision and recall. The potential impact on blood supply was estimated by predicting deferral at earlier or later donation dates.ResultsWe present a model that predicts Hb deferral in an explainable way. If used in practice, 64% of non-deferred donors would be invited on or before their original donation date, while 80% of deferred donors would be invited later.ConclusionBy using this model to invite donors, the number of blood bank visits would increase by 15%, while deferral rates would decrease by 60% (currently 3% for women and 1% for men).
| S-EPMC9826045 | biostudies-literature
Project description:In recent years, a number of machine learning models for the prediction of the skin sensitization potential of small organic molecules have been reported and become available. These models generally perform well within their applicability domains but, as a result of the use of molecular fingerprints and other non-intuitive descriptors, the interpretability of the existing models is limited. The aim of this work is to develop a strategy to replace the non-intuitive features by predicted outcomes of bioassays. We show that such replacement is indeed possible and that as few as ten interpretable, predicted bioactivities are sufficient to reach competitive performance. On a holdout data set of 257 compounds, the best model ("Skin Doctor CP:Bio") obtained an efficiency of 0.82 and an MCC of 0.52 (at the significance level of 0.20). Skin Doctor CP:Bio is available free of charge for academic research. The modeling strategies explored in this work are easily transferable and could be adopted for the development of more interpretable machine learning models for the prediction of the bioactivity and toxicity of small organic compounds.
| S-EPMC8402010 | biostudies-literature