Dr. Chindelevitch’s research programme focuses on the mathematical and computational modeling of antimicrobial resistance in infectious disease, both on the molecular level (using computational and systems biology) as well as on the population level (using epidemiology and population genetics). He is also interested in the application of science to policy towards the goal of improving patient outcomes, especially in low-income or low-resource settings. His other research interests focus on computational biology, algorithm development, discrete optimization, and machine learning.
Drug resistance prediction
Traditionally carried out through costly and time-consuming methods such as drug sensitivity testing, the prediction of drug resistance in bacterial pathogens can now be performed through the analysis of whole-genome sequence data. The challenge is to carry out this prediction in a way that is both accurate (producing few incorrect predictions) and interpretable (identifying specific mutations that drive drug resistance). Dr. Chindelevitch’s recent work is focused on addressing this challenge, with methods involving exact optimization and modular deep neural networks.
Epidemiological relatedness from whole-genome sequences
Another important use of whole-genome sequences is to identify the patients in an outbreak that may be related by a transmission chain. Traditionally, such an analysis may have been carried through a genotyping method that examines a specific set of genes or counts the copy numbers of tandem repeats. Recent developments have enabled this to be done in a more precise manner by examining the entire genome. Dr. Chindelevitch’s work integrates these methods into a unified platform to rapidly identify epidemiologically related samples in an outbreak.
Control of the spread of drug-resistant tuberculosis
As part of his postdoctoral research Dr. Chindelevitch developed a compartmental model that accounts for the joint natural history of tuberculosis and HIV, with a particular emphasis on drug-resistant tuberculosis. This model was calibrated to data from South Africa and used to investigate the effects of various interventions aimed at reducing the burden of these diseases in order to help policy-makers select the most impactful and cost-effective alternative.
Dr. Chindelevitch holds a PhD in Applied Mathematics from MIT and a BSc in Mathematics and Computer Science from McGill University. His doctoral work focused on metabolic models of tuberculosis (TB). Dr. Chindelevitch completed a postdoctoral fellowship in Ted Cohen's group from 2012 to 2015 and was a faculty member at the School of Computing Science at Simon Fraser University from 2015 to 2020, where he received an Alfred P. Sloan Research Fellowship in Computational and Evolutionary Molecular Biology. He joined the Department of Infectious Disease at Imperial College in 2020.
Dr. Chindelevitch also has experience of working in industry and non-profit organizations. He worked as a computational biologist at Pfizer between his PhD and his postdoctoral fellowship, where he developed methods for interpreting gene expression, genetic and metabolomic data using a large network of known biological relationships. He has also worked at the Massachusetts General Hospital and the Clinton Health Access Initiative and is currently consulting the Foundation for Innovative New Diagnostics on drug resistance in tuberculosis.
et al., 2023, Generalizations of the genomic rank distance to indels., Bioinformatics, Vol:39
Xue Z, Chindelevitch L, Guichard F, 2023, Supply-driven evolution: mutation bias and trait-fitness distributions can drive macro-evolutionary dynamics, Frontiers in Ecology and Evolution, Vol:10, ISSN:2296-701X
et al., 2022, Deep clustering of bacterial tree images, Philosophical Transactions of the Royal Society B-biological Sciences, Vol:377, ISSN:0962-8436
et al., 2022, The 2021 WHO catalogue of Mycobacterium tuberculosis complex mutations associated with drug resistance: a genotypic analysis., The Lancet Microbe, Vol:3, ISSN:2666-5247, Pages:e265-e273
et al., 2022, Aristotle: stratified causal discovery for omics data, Bmc Bioinformatics, Vol:23, ISSN:1471-2105