Dear All,
I have been conducting analyses on a dataset of bacterial pathogens which included several relevant biases: imbalance in host association, temporal distribution (which hosts have been sampled when and how frequently) and geographic origin. The preprint including a visualisation of these biases can be found here: https://doi.org/10.64898/2026.01.06.697867
I have used both discrete trait analysis in BEAST1 with a fixed topology and traits modelled on top as well as MASCOT v3, expecting the latter to be more robust towards the host sampling bias.
We have concluded that due to the combination of these biases, inference of the root- and early node-host association cannot be be reported with certainty. However, in a recent discussion there has been disagreement over whether the direction of migration rates (whether transitions from host A->B or B->A happened more frequently over time) is affected in the same way. In short, if we say we do not trust the inferred root association, due to the combination of host but also temporal sampling bias, would one be cautious of the migration rates, too?
Beyond the original MASCOT publications, I have consulted multiple articles on this (see below). I had personally drawn the conclusion that while encountering the same results with high support in both types of analyses gives some confidence, I reserve a level of caution and would still report the definitive direction of migration as unresolved.
Does someone from the community with MASCOT experience have an opinion on this?
Thank you very much!
Antonia
Layan M, Müller NF, Dellicour S, De Maio N, Bourhy H, Cauchemez S, et al. Impact and mitigation of sampling bias to determine viral spread: Evaluating discrete phylogeography through CTMC modeling and structured coalescent model approximations. Virus Evol. 2023;9(1):vead010.
Hall MD, Woolhouse ME, Rambaut A. The effects of sampling strategy on the quality of reconstruction of viral population dynamics using Bayesian skyline family coalescent methods: A simulation study. Virus Evol. 2016;2(1):vew003.
. De Maio N, Wu CH, O'Reilly KM, Wilson D. New Routes to Phylogeography: A Bayesian Structured Coalescent Approximation. PLoS Genet. 2015;11(8):e1005421.
Kalkauskas A, Perron U, Sun Y, Goldman N, Baele G, Guindon S, et al. Sampling bias and model choice in continuous phylogeography: Getting lost on a random walk. PLoS Comput Biol. 2021;17(1):e1008561.
Bloomfield S, Vaughan T, Benschop J, Marshall J, Hayman D, Biggs P, et al. Investigation of the validity of two Bayesian ancestral state reconstruction models for estimating Salmonella transmission during outbreaks. PLoS One. 2019;14(7):e0214169.
