Dataset Information

Multi-Label Random Forest Model for Tuberculosis Drug Resistance Classification and Mutation Ranking.

ABSTRACT: Resistance prediction and mutation ranking are important tasks in the analysis of Tuberculosis sequence data. Due to standard regimens for the use of first-line antibiotics, resistance co-occurrence, in which samples are resistant to multiple drugs, is common. Analysing all drugs simultaneously should therefore enable patterns reflecting resistance co-occurrence to be exploited for resistance prediction. Here, multi-label random forest (MLRF) models are compared with single-label random forest (SLRF) for both predicting phenotypic resistance from whole genome sequences and identifying important mutations for better prediction of four first-line drugs in a dataset of 13402 Mycobacterium tuberculosis isolates. Results confirmed that MLRFs can improve performance compared to conventional clinical methods (by 18.10%) and SLRFs (by 0.91%). In addition, we identified a list of candidate mutations that are important for resistance prediction or that are related to resistance co-occurrence. Moreover, we found that retraining our analysis to a subset of top-ranked mutations was sufficient to achieve satisfactory performance. The source code can be found at http://www.robots.ox.ac.uk/~davidc/code.php.

SUBMITTER: Kouchaki S

PROVIDER: S-EPMC7188832 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Multi-Label Random Forest Model for Tuberculosis Drug Resistance Classification and Mutation Ranking.

Kouchaki Samaneh S Yang Yang Y Lachapelle Alexander A Walker Timothy M TM Walker A Sarah AS Peto Timothy E A TEA Crook Derrick W DW Clifton David A DA

Frontiers in microbiology 20200422

Resistance prediction and mutation ranking are important tasks in the analysis of Tuberculosis sequence data. Due to standard regimens for the use of first-line antibiotics, resistance co-occurrence, in which samples are resistant to multiple drugs, is common. Analysing all drugs simultaneously should therefore enable patterns reflecting resistance co-occurrence to be exploited for resistance prediction. Here, multi-label random forest (MLRF) models are compared with single-label random forest ( ...[more]

PMID: 32390972

Similar Datasets

Project description:Neurodegenerative disorders, such as Alzheimer's disease, are associated with changes in multiple neuroimaging and biological measures. These may provide complementary information for diagnosis and prognosis. We present a multi-modality classification framework in which manifolds are constructed based on pairwise similarity measures derived from random forest classifiers. Similarities from multiple modalities are combined to generate an embedding that simultaneously encodes information about all the available features. Multi-modality classification is then performed using coordinates from this joint embedding. We evaluate the proposed framework by application to neuroimaging and biological data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Features include regional MRI volumes, voxel-based FDG-PET signal intensities, CSF biomarker measures, and categorical genetic information. Classification based on the joint embedding constructed using information from all four modalities out-performs the classification based on any individual modality for comparisons between Alzheimer's disease patients and healthy controls, as well as between mild cognitive impairment patients and healthy controls. Based on the joint embedding, we achieve classification accuracies of 89% between Alzheimer's disease patients and healthy controls, and 75% between mild cognitive impairment patients and healthy controls. These results are comparable with those reported in other recent studies using multi-kernel learning. Random forests provide consistent pairwise similarity measures for multiple modalities, thus facilitating the combination of different types of feature data. We demonstrate this by application to data in which the number of features differs by several orders of magnitude between modalities. Random forest classifiers extend naturally to multi-class problems, and the framework described here could be applied to distinguish between multiple patient groups in the future.

Project description:"Drug resistance is an unavoidable consequence of the use of drugs; however, the emergence of multi-drug resistance can be managed by accurate diagnosis and tailor-made regimens."Antimicrobial resistance (AMR), is one of the most paramount health perils that has emerged in the 21st century. The global increase in drug-resistant strains of various bacterial pathogens prompted the World Health Organization (WHO) to develop a priority list of AMR pathogens. Mycobacterium tuberculosis (Mtb), an acid-fast bacillus that causes tuberculosis (TB), merits being one of the highest priority pathogens on this list since drug-resistant TB (DR-TB) accounts for ∼29% of deaths attributable to AMR. In recent years, funded collaborative efforts of researchers from academia, not-for-profit virtual R&D organizations and industry have resulted in the continuous growth of the TB drug discovery and development pipeline. This has so far led to the accelerated regulatory approval of bedaquiline and delamanid for the treatment of DR-TB. However, despite the availability of drug regimes, the current cure rate for multi-drug-resistant TB (MDR-TB) and extensively drug-resistant TB (XDR-TB) treatment regimens is 50% and 30%, respectively. It is to be noted that these regimens are administered over a long duration and have a serious side effect profile. Coupled with poor patient adherence, this has led to further acquisition of drug resistance and treatment failure. There is therefore an urgent need to develop new TB drugs with novel mechanism of actions (MoAs) and associated regimens.This Account recapitulates drug resistance in TB, existing challenges in addressing DR-TB, new drugs and regimens in development, and potential ways to treat DR-TB. We highlight our research aimed at identifying novel small molecule leads and associated targets against TB toward contributing to the global TB drug discovery and development pipeline. Our work mainly involves screening of various small molecule chemical libraries in phenotypic whole-cell based assays to identify hits for medicinal chemistry optimization, with attendant deconvolution of the MoA. We discuss the identification of small molecule chemotypes active against Mtb and subsequent structure-activity relationships (SAR) and MoA deconvolution studies. This is followed by a discussion on a chemical series identified by whole-cell cross-screening against Mtb, for which MoA deconvolution studies revealed a pathway that explained the lack of in vivo efficacy in a mouse model of TB and reiterated the importance of selecting an appropriate growth medium during phenotypic screening. We also discuss our efforts on drug repositioning toward addressing DR-TB. In the concluding section, we preview some promising future directions and the challenges inherent in advancing the drug pipeline to address DR-TB.

Dataset Information

Multi-Label Random Forest Model for Tuberculosis Drug Resistance Classification and Mutation Ranking.

Publications

Multi-Label Random Forest Model for Tuberculosis Drug Resistance Classification and Mutation Ranking.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets