Large-scale genomic analysis shows association between homoplastic genetic variation in <italic>Mycobacterium tuberculosis</italic> genes and meningeal or pulmonary tuberculosis.

Background: Meningitis is the most severe manifestation of tuberculosis. It is largely unknown why some people develop pulmonary TB (PTB) and others TB meningitis (TBM); we examined if the genetic background of infecting <italic>M. tuberculosis</italic> strains may be relevant. Methods: We whole-genome sequenced <italic>M. tuberculosis</italic> strains isolated from 322 HIV-negative tuberculosis patients from Indonesia and compared isolates from patients with TBM (<italic>n</italic> = 106) and PTB (<italic>n</italic> = 216). Using a phylogeny-adjusted genome-wide association method to count homoplasy events we examined phenotype-related changes at specific loci or genes in parallel branches of the phylogenetic tree. Enrichment scores for the TB phenotype were calculated on single nucleotide polymorphism (SNP), gene, and pathway level. Genetic associations were validated in an independent set of isolates. Results: Strains belonged to the East-Asian lineage (36.0%), Euro-American lineage (61.5%), and Indo-Oceanic lineage (2.5%). We found no association between lineage and phenotype (Chi-square = 4.556; <italic>p</italic> = 0.207). Large genomic differences were observed between isolates; the minimum pairwise genetic distance varied from 17 to 689 SNPs. Using the phylogenetic tree, based on 28,544 common variable positions, we selected 54 TBM and 54 PTB isolates in terminal branch sets with distinct phenotypes. Genetic variation in Rv0218, and absence of Rv3343c, and <italic>nanK</italic> were significantly associated with disease phenotype in these terminal branch sets, and confirmed in the validation set of 214 unpaired isolates. Conclusions: Using homoplasy counting we identified genetic variation in three separate genes to be associated with the TB phenotype, including one (Rv0218) which encodes a secreted protein that could play a role in host-pathogen interaction by altering pathogen recognition or acting as virulence effector. [ABSTRACT FROM AUTHOR]