Genomic prediction of tuberculosis drug-resistance: benchmarking existing databases and prediction algorithms.
Ontology highlight
ABSTRACT:
SUBMITTER: Ngo TM
PROVIDER: S-EPMC6368788 | biostudies-literature | 2019 Feb
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:BackgroundThe increasing prevalence of bovine tuberculosis (bTB) in the UK and the limitations of the currently available diagnostic and control methods require the development of complementary approaches to assist in the sustainable control of the disease. One potential approach is the identification of animals that are genetically more resistant to bTB, to enable breeding of animals with enhanced resistance. This paper focuses on prediction of resistance to bTB. We explore estimation of direct genomic estimated breeding values (DGVs) for bTB resistance in UK dairy cattle, using dense SNP chip data, and test these genomic predictions for situations when disease phenotypes are not available on selection candidates.Methodology/principal findingsWe estimated DGVs using genomic best linear unbiased prediction methodology, and assessed their predictive accuracies with a cross validation procedure and receiver operator characteristic (ROC) curves. Furthermore, these results were compared with theoretical expectations for prediction accuracy and area-under-the-ROC-curve (AUC). The dataset comprised 1151 Holstein-Friesian cows (bTB cases or controls). All individuals (592 cases and 559 controls) were genotyped for 727,252 loci (Illumina Bead Chip). The estimated observed heritability of bTB resistance was 0.23±0.06 (0.34 on the liability scale) and five-fold cross validation, replicated six times, provided a prediction accuracy of 0.33 (95% C.I.: 0.26, 0.40). ROC curves, and the resulting AUC, gave a probability of 0.58, averaged across six replicates, of correctly classifying cows as diseased or as healthy based on SNP chip genotype alone using these data.Conclusions/significanceThese results provide a first step in the investigation of the potential feasibility of genomic selection for bTB resistance using SNP data. Specifically, they demonstrate that genomic selection is possible, even in populations with no pedigree data and on animals lacking bTB phenotypes. However, a larger training population will be required to improve prediction accuracies.
| S-EPMC4014548 | biostudies-literature
Project description:Effective and timely antibiotic treatment depends on accurate and rapid in silico antimicrobial-resistant (AMR) predictions. Existing statistical rule-based Mycobacterium tuberculosis (MTB) drug resistance prediction methods using bacterial genomic sequencing data often achieve varying results: high accuracy on some antibiotics but relatively low accuracy on others. Traditional machine learning (ML) approaches have been applied to classify drug resistance for MTB and have shown more stable performance. However, there is no study that uses deep learning architecture like Convolutional Neural Network (CNN) on a large and diverse cohort of MTB samples for AMR prediction. We developed 24 binary classifiers of MTB drug resistance status across eight anti-MTB drugs and three different ML algorithms: logistic regression, random forest and 1D CNN using a training dataset of 10,575 MTB isolates collected from 16 countries across six continents, where an extended pan-genome reference was used for detecting genetic features. Our 1D CNN architecture was designed to integrate both sequential and non-sequential features. In terms of F1-scores, 1D CNN models are our best classifiers that are also more accurate and stable than the state-of-the-art rule-based tool Mykrobe predictor (81.1 to 93.8%, 93.7 to 96.2%, 93.1 to 94.8%, 95.9 to 97.2% and 97.1 to 98.2% for ethambutol, rifampicin, pyrazinamide, isoniazid and ofloxacin respectively). We applied filter-based feature selection to find AMR relevant features. All selected variant features are AMR-related ones in CARD database. 78.8% of them are also in the catalogue of MTB mutations that were recently identified as drug resistance-associated ones by WHO. To facilitate ML model development for AMR prediction, we packaged every step into an automated pipeline and shared the source code at https://github.com/KuangXY3/MTB-AMR-classification-CNN .
| S-EPMC8844416 | biostudies-literature
Project description:BackgroundVariants that disrupt mRNA splicing account for a sizable fraction of the pathogenic burden in many genetic disorders, but identifying splice-disruptive variants (SDVs) beyond the essential splice site dinucleotides remains difficult. Computational predictors are often discordant, compounding the challenge of variant interpretation. Because they are primarily validated using clinical variant sets heavily biased to known canonical splice site mutations, it remains unclear how well their performance generalizes.ResultsWe benchmarked eight widely used splicing effect prediction algorithms, leveraging massively parallel splicing assays (MPSAs) as a source of experimentally determined ground-truth. MPSAs simultaneously assay many variants to nominate candidate SDVs. We compared experimentally measured splicing outcomes with bioinformatic predictions for 3,616 variants in five genes. Algorithms' concordance with MPSA measurements, and with each other, was lower for exonic than intronic variants, underscoring the difficulty of identifying missense or synonymous SDVs. Deep learning-based predictors trained on gene model annotations achieved the best overall performance at distinguishing disruptive and neutral variants. Controlling for overall call rate genome-wide, SpliceAI and Pangolin also showed superior overall sensitivity for identifying SDVs. Finally, our results highlight two practical considerations when scoring variants genome-wide: finding an optimal score cutoff, and the substantial variability introduced by differences in gene model annotation, and we suggest strategies for optimal splice effect prediction in the face of these issues.ConclusionSpliceAI and Pangolin showed the best overall performance among predictors tested, however, improvements in splice effect prediction are still needed especially within exons.
| S-EPMC10187268 | biostudies-literature
Project description:BackgroundVariants that disrupt mRNA splicing account for a sizable fraction of the pathogenic burden in many genetic disorders, but identifying splice-disruptive variants (SDVs) beyond the essential splice site dinucleotides remains difficult. Computational predictors are often discordant, compounding the challenge of variant interpretation. Because they are primarily validated using clinical variant sets heavily biased to known canonical splice site mutations, it remains unclear how well their performance generalizes.ResultsWe benchmark eight widely used splicing effect prediction algorithms, leveraging massively parallel splicing assays (MPSAs) as a source of experimentally determined ground-truth. MPSAs simultaneously assay many variants to nominate candidate SDVs. We compare experimentally measured splicing outcomes with bioinformatic predictions for 3,616 variants in five genes. Algorithms' concordance with MPSA measurements, and with each other, is lower for exonic than intronic variants, underscoring the difficulty of identifying missense or synonymous SDVs. Deep learning-based predictors trained on gene model annotations achieve the best overall performance at distinguishing disruptive and neutral variants, and controlling for overall call rate genome-wide, SpliceAI and Pangolin have superior sensitivity. Finally, our results highlight two practical considerations when scoring variants genome-wide: finding an optimal score cutoff, and the substantial variability introduced by differences in gene model annotation, and we suggest strategies for optimal splice effect prediction in the face of these issues.ConclusionSpliceAI and Pangolin show the best overall performance among predictors tested, however, improvements in splice effect prediction are still needed especially within exons.
| S-EPMC10734170 | biostudies-literature
Project description:Pearson's correlation coefficient (ρ) is the most commonly reported metric of the success of prediction in genomic selection (GS). However, in real breeding ρ may not be very useful for assessing the quality of the regression in the tails of the distribution, where individuals are chosen for selection. This research used 14 maize and 16 wheat data sets with different trait-environment combinations. Six different models were evaluated by means of a cross-validation scheme (50 random partitions each, with 90% of the individuals in the training set and 10% in the testing set). The predictive accuracy of these algorithms for selecting individuals belonging to the best α=10, 15, 20, 25, 30, 35, 40% of the distribution was estimated using Cohen's kappa coefficient (κ) and an ad hoc measure, which we call relative efficiency (RE), which indicates the expected genetic gain due to selection when individuals are selected based on GS exclusively. We put special emphasis on the analysis for α=15%, because it is a percentile commonly used in plant breeding programmes (for example, at CIMMYT). We also used ρ as a criterion for overall success. The algorithms used were: Bayesian LASSO (BL), Ridge Regression (RR), Reproducing Kernel Hilbert Spaces (RHKS), Random Forest Regression (RFR), and Support Vector Regression (SVR) with linear (lin) and Gaussian kernels (rbf). The performance of regression methods for selecting the best individuals was compared with that of three supervised classification algorithms: Random Forest Classification (RFC) and Support Vector Classification (SVC) with linear (lin) and Gaussian (rbf) kernels. Classification methods were evaluated using the same cross-validation scheme but with the response vector of the original training sets dichotomised using a given threshold. For α=15%, SVC-lin presented the highest κ coefficients in 13 of the 14 maize data sets, with best values ranging from 0.131 to 0.722 (statistically significant in 9 data sets) and the best RE in the same 13 data sets, with values ranging from 0.393 to 0.948 (statistically significant in 12 data sets). RR produced the best mean for both κ and RE in one data set (0.148 and 0.381, respectively). Regarding the wheat data sets, SVC-lin presented the best κ in 12 of the 16 data sets, with outcomes ranging from 0.280 to 0.580 (statistically significant in 4 data sets) and the best RE in 9 data sets ranging from 0.484 to 0.821 (statistically significant in 5 data sets). SVC-rbf (0.235), RR (0.265) and RHKS (0.422) gave the best κ in one data set each, while RHKS and BL tied for the last one (0.234). Finally, BL presented the best RE in two data sets (0.738 and 0.750), RFR (0.636) and SVC-rbf (0.617) in one and RHKS in the remaining three (0.502, 0.458 and 0.586). The difference between the performance of SVC-lin and that of the rest of the models was not so pronounced at higher percentiles of the distribution. The behaviour of regression and classification algorithms varied markedly when selection was done at different thresholds, that is, κ and RE for each algorithm depended strongly on the selection percentile. Based on the results, we propose classification method as a promising alternative for GS in plant breeding.
| S-EPMC4023444 | biostudies-literature
Project description:In clinical treatment, two or more drugs (i.e. drug combination) are simultaneously or successively used for therapy with the purpose of primarily enhancing the therapeutic efficacy or reducing drug side effects. However, inappropriate drug combination may not only fail to improve efficacy, but even lead to adverse reactions. Therefore, according to the basic principle of improving the efficacy and/or reducing adverse reactions, we should study drug-drug interactions (DDIs) comprehensively and thoroughly so as to reasonably use drug combination. In this review, we first introduced the basic conception and classification of DDIs. Further, some important publicly available databases and web servers about experimentally verified or predicted DDIs were briefly described. As an effective auxiliary tool, computational models for predicting DDIs can not only save the cost of biological experiments, but also provide relevant guidance for combination therapy to some extent. Therefore, we summarized three types of prediction models (including traditional machine learning-based models, deep learning-based models and score function-based models) proposed during recent years and discussed the advantages as well as limitations of them. Besides, we pointed out the problems that need to be solved in the future research of DDIs prediction and provided corresponding suggestions.
| S-EPMC10782925 | biostudies-literature
Project description:Whole-genome sequencing allows rapid detection of drug-resistant Mycobacterium tuberculosis isolates. However, the availability of high-quality data linking quantitative phenotypic drug susceptibility testing (DST) and genomic data have thus far been limited. We determined drug resistance profiles of 176 genetically diverse clinical M. tuberculosis isolates from the Democratic Republic of the Congo, Ivory Coast, Peru, Thailand, and Switzerland by quantitative phenotypic DST for 11 antituberculous drugs using the BD Bactec MGIT 960 system and 7H10 agar dilution to generate a cross-validated phenotypic DST readout. We compared DST results with predicted drug resistance profiles inferred by whole-genome sequencing. Classification of strains by the two phenotypic DST methods into resistotype/wild-type populations was concordant in 73 to 99% of cases, depending on the drug. Our data suggest that the established critical concentration (5?mg/liter) for ethambutol resistance (MGIT 960 system) is too high and misclassifies strains as susceptible, unlike 7H10 agar dilution. Increased minimal inhibitory concentrations were explained by mutations identified by whole-genome sequencing. Using whole-genome sequences, we were able to predict quantitative drug resistance levels for the majority of drug resistance mutations. Predicting quantitative levels of drug resistance by whole-genome sequencing was partially limited due to incompletely understood drug resistance mechanisms. The overall sensitivity and specificity of whole-genome-based DST were 86.8% and 94.5%, respectively. Despite some limitations, whole-genome sequencing has the potential to infer resistance profiles without the need for time-consuming phenotypic methods.
| S-EPMC6496161 | biostudies-literature
Project description:Tuberculosis is a global pandemic disease with a rising burden of antimicrobial resistance. As a result, the World Health Organization (WHO) has a goal of enabling universal access to drug susceptibility testing (DST). Given the slowness of and infrastructure requirements for phenotypic DST, whole-genome sequencing, followed by genotype-based prediction of DST, now provides a route to achieving this. Since a central component of genotypic DST is to detect the presence of any known resistance-causing mutations, a natural approach is to use a reference graph that allows encoding of known variation. We have developed DrPRG (Drug resistance Prediction with Reference Graphs) using the bacterial reference graph method Pandora. First, we outline the construction of a Mycobacterium tuberculosis drug resistance reference graph. The graph is built from a global dataset of isolates with varying drug susceptibility profiles, thus capturing common and rare resistance- and susceptible-associated haplotypes. We benchmark DrPRG against the existing graph-based tool Mykrobe and the haplotype-based approach of TBProfiler using 44 709 and 138 publicly available Illumina and Nanopore samples with associated phenotypes. We find that DrPRG has significantly improved sensitivity and specificity for some drugs compared to these tools, with no significant decreases. It uses significantly less computational memory than both tools, and provides significantly faster runtimes, except when runtime is compared to Mykrobe with Nanopore data. We discover and discuss novel insights into resistance-conferring variation for M. tuberculosis - including deletion of genes katG and pncA - and suggest mutations that may warrant reclassification as associated with resistance.
| S-EPMC10483414 | biostudies-literature
Project description:Mycobacterium tuberculosis has acquired drug resistance to all drugs that have been used against it, including those only recently introduced into clinical practice. Compared to other bacteria, it has a well conserved genome due to its role as an obligate human pathogen that has adapted to a niche over five to ten thousand years. These features facilitate reconstruction and dating of M. tuberculosis phylogenies, giving key insights into how resistance has been acquired and spread globally. Resistance to each new drug has occurred within five to ten years of clinical use and has occurred even more rapidly with recently introduced drugs. In most cases, resistance-conferring mutations come with a fitness cost, but this can be overcome by compensatory mutations which restore fitness to that of wild-type bacteria. It is likely that M. tuberculosis acquires drug resistance while maintaining limited genomic variability due the generation of low frequency within-host variation, combined with ongoing purifying selection causing loss of variants without a clear fitness advantage. However, variants that do confer an advantage, such as drug resistance, can increase in prevalence amongst all bacteria within a host and become the dominant clone. These resistant strains can then be transmitted leading to primary drug resistant infection in a new host. As many countries move towards genomic methods for diagnosis of M. tuberculosis infection and drug resistance, it is important to be aware of the implications for the evolution of resistance. Currently, understanding of resistance-conferring mutations is incomplete, and some targeted genetic diagnostics create their own selective pressures. We discuss an example where a rifampicin resistance-conferring mutation which was not routinely covered by standard testing became dominant. Finally, resistance to new drugs such as bedaquiline and delamanid is caused by individually rare mutations occurring across a large mutational genomic target that have been detected over a short time, and do not provide statistical power for genotype-phenotype correlation - in contrast to longer-established drugs that form the backbone of drug-sensitive antituberculosis therapy. Therefore, we need a different approach to identify resistance-conferring mutations of new drugs before their resistance becomes widespread, abrogating their usefulness.
| S-EPMC9585206 | biostudies-literature
Project description:This study aimed to identify small molecules that inhibit vascular calcification with computational approach. To examine the mechanism of calcification inhibition, proteomics analysis was performed using aorta tissue samples to clarify therapeutic target.
2022-02-17 | PXD029963 | Pride