Minos: variant adjudication and joint genotyping of cohorts of bacterial genomes

Abstract

There are many short-read variant-calling tools, with different strengths and weaknesses. We present a tool, Minos, which combines outputs from arbitrary variant callers, increasing recall without loss of precision. We benchmark on 62 samples from three bacterial species and an outbreak of 385 Mycobacterium tuberculosis samples. Minos also enables joint genotyping; we demonstrate on a large (N=13k) M. tuberculosis cohort, building a map of non-synonymous SNPs and indels in a region where all such variants are assumed to cause rifampicin resistance. We quantify the correlation with phenotypic resistance and then replicate in a second cohort (N=10k).

Genomic Sequencing from Sputum for Tuberculosis Disease Diagnosis, Lineage Determination, and Drug Susceptibility Prediction

“…The composition of each isolate’s sequencing data was determined by matching each read’s best mapping in the contamination alignment file to a classification category of M. tuberculosis , nontuberculous mycobacteria (NTM), other bacteria, virus, human, or unmapped. Mykrobe (v.0.10.0) was used to assign lineage and predict drug susceptibility using parameters described in Hall et al ( 10 , 11 , 15 ). The complete list of gene targets and the resistance conferring mutation catalog used by Mykrobe was previously reported and is available online and in Tables S1 and S2 in the supplemental material ( 10 , 16 ).…”

Epidemiological cut-off values for a 96-well broth microdilution plate for high-throughput research antibiotic susceptibility testing ofM. tuberculosis

“…These calls were then passed to the adjudication software minos, which produces a graph representation of the reference genome plus conflicting calls from the two callsets, and then remaps reads to the graph to adjudicate statistically. This adjudication process, and the performance of the combined samtools/cortex callset, are documented (53). All of this process, including versions of samtools and cortex and the reference genomes for filtering) is encapsulated in Clockwork version 0.8.3 (54).…”

Repeated evolution of bedaquiline resistance inMycobacterium tuberculosisis driven by truncation ofmmpR5

“…Per-sample variant calling was completed using Clockwork v0.10.0 (https://github.com/iqbal-lab-org/clockwork) against the reference genome H37Rv (accession: NC_000962.3). Clockwork is a variant caller which runs two other independent callers (SAMtools 22 , Cortex 23 ) on each isolate, and then builds a graph genome of the (reference genome and) two sets of variant calls, and remaps the reads to this graph, in order to adjudicate between the two callers where they disagree 24 . Having done this, Minos 24 v0.12.0 was used to collect a non-redundant list of all segregating variants in the cohort, harmonise variant representation in positions where indels overlap SNPs (or other indels), and then remap the decontaminated reads one more time, to genotype all samples at all variants, providing consistent VCFs for all samples.…”

“…(accession: NC_000962.3). Clockwork is a variant caller which runs two other independent callers (SAMtools 22 , Cortex 23 ) on each isolate, and then builds a graph genome of the (reference genome and) two sets of variant calls, and remaps the reads to this graph, in order to adjudicate between the two callers where they disagree 24 .…”