The molecular clock of Mycobacterium tuberculosis

Abstract

23The molecular clock and its phylogenetic applications to genomic data have changed how we study and 24 understand one of the major human pathogens, Mycobacterium tuberculosis (MTB), the causal agent of 25 tuberculosis. Genome sequences of MTB strains sampled at different times are increasingly used to 26 infer when a particular outbreak begun, when a drug resistant clone appeared and expanded, or when a 27 strain was introduced into a specific region. Despite the growing importance of the molecular clock in 28 tuberculosis research, there is a lack of consensus as to whether MTB displays a clocklike behavior and 29 about its rate of evolution. Here we performed a systematic study of the MTB molecular clock on a 30 large genomic data set (6,285 strains), covering most of the global MTB diversity and representing 31 different epidemiological settings. We found wide variation in the degree of clocklike structure among 32 data sets, indicating that sampling times are sometimes insufficient to calibrate the clock of MTB. For 33 data sets with temporal structure, we found that MTB genomes accumulate between 1x10 -8 and 5x10 -7 34 nucleotide changes per-site-per-year, which corresponds to 0.04 -2.2 SNPs per-genome- per-year. 35 Contrary to what expected, these estimates were not dependent on the time of the calibration points as 36 they did not change significantly when we used epidemiological isolates (sampled in the last 40 years) 37 or ancient DNA samples (about 1,000 years old) to calibrate the tree. Additionally, the uncertainty and 38 the discrepancies in the results of different methods were often large, highlighting the importance of 39 using different methods, and of considering carefully their assumptions and limitations. 40 41 Keywords 42 Evolution, Phylogenetics, Pathogen, Bacteria 43 44 45 3 Significance Statement 46 47One of the major recent advancement in evolutionary biology is the development of statistical methods 48 to infer the past evolutionary history of species and populations with genomic data. In the last five 49 years, many researchers have used the molecular clock to study the evolution of Mycobacterium 50 tuberculosis, a bacterial pathogen that causes tuberculosis and is responsible for millions of human 51deaths every year. The application of the molecular clock to tuberculosis is extremely useful to 52 understand the evolution of drug resistance, the spread of different strains and the origin of the disease. 53Since some of these studies found contrasting results, we performed a systematic analysis of the 54 molecular clock of MTB. This study will provide an important guideline for future analyses of 55 tuberculosis and other organisms.

An African origin forMycobacterium bovis

“…The tip-dating calibration provided accurate results for the emergency of BCG strains (TreeS2.txt), indicating that the method can reliably infer divergence times at least for events occurring in the last 100 years. However, extrapolating the clock rate to date older divergence events should be done with caution [37].…”

“…For the molecular clock analyses we considered only genomes for which the date of isolation was known (n=2,058). We used a pipeline similar to that reported in [37]. We built SNPs alignments including variable positions with less than 10% of missing data.…”

A sister lineage of theMycobacterium tuberculosis complexdiscovered in the African Great Lakes region

“…Isoniazid and rifampicin were introduced in TB treatments in Rwanda and Uganda in the late fifties and early nineties, respectively. Therefore, the ~100 SNPs distance separating these two strains from their MRCA would imply a rapid molecular clock for L8, above the upper limit of 2.2 SNPs/genome/year most recently estimated for other MTBC clades 38 . However, this mutation rate cannot be confirmed until additional L8 samples are uncovered.…”

“…The maximum likelihood phylogeny was inferred with RAxML v.8.2.8 60 using an alignment containing only polymorphic sites and the branch lengths of the tree were rescaled using invariant sites (rescaled_branch_length = (branch_length * alignment_length) / (alignment_length +invariant_sites)) 38,61 .…”

Multiple merger genealogies in outbreaks ofMycobacterium tuberculosis

“…These results provide compelling evidence that, in most cases, MMC models have a better fit to data from MTB outbreaks compared to standard models based on In the light of our results, it is not surprising that most of these studies found evidence for population growth, as it is known that performing demographic inference on MMC genealogies assuming the Kingman coalescent fits high growth rates (Eldon et al 2015). Moreover, a recent study fitted an exponential growth (or shrinkage) model to 21 MTB data sets, and found evidence for population growth for 14 of them, for seven data sets it was not possible to reject the hypothesis of constant population size, and none of them resulted in population shrinkage (Menardo et al 2019). One additional factor that could bias the results of Kingman-based demographic inference towards population growth is sampling bias (Lapierre et al 2016).…”