This is an archived static version of the original phylobabble.org discussion site.

Quantifying changes in clonal populations (phylo-epi)

adamr

Dear Phylobabble community. I’m working with epidemiological surveillance data for a bacterial pathogen (N. meningitidis) and am trying to figure out how to describe the changes in the population of disease-causing isolates over time. Conventionally, this has been done by placing the bacterial isolates into discrete categories (e.g. by MLST) and then reporting the abundance of the various categories, and changes in abundance. I’m trying to obtain similar information using phylogenies inferred from whole genome sequence data. Can you suggest some statistics that would help us to quantify how quickly the population is changing and good ways to identify clades that are expanding or contracting exceptionally fast? If so, I’d greatly appreciate hearing about them.

Thanks Adam

cmeehan

Hi Adam,

If I understand you correctly, you are trying to estimate population sizes and changes over time, yes? There are a couple of ways to do this, primarily within Bayesian frameworks. Firstly you could use something like hierBAPS to look at genomic clusters and see if specific clusters are larger over time than others. The best approach though is to undertake a Bayesian skyline analysis within BEAST or BEAST2 for each of your populations. This will allow you to estimate the changes in population sizes and see if one is expanding/contracting differently from the others.

Cheers, Con

adamr

Thanks for the suggestion Con. I’ll look into those approaches. Regards, Adam

adamr

Update. This paper (Volz and Didelot) addresses the question I was asking about…at least, it addresses the part about identify clades that are rapidly expanding or contracting: https://www.biorxiv.org/content/10.1101/704528v1