Comparing different computational approaches for detecting long‐term vertical transmission in host‐associated microbiota - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Molecular Ecology Année : 2022

Comparing different computational approaches for detecting long‐term vertical transmission in host‐associated microbiota

Résumé

Long-term vertical transmissions of gut bacteria are thought to be frequent and functionally important in mammals. Several phylogenetic-based approaches have been proposed to detect, among species-rich microbiota, the bacteria that have been vertically transmitted during a host clade radiation. Applied to mammal microbiota, these methods have sometimes led to conflicting results; in addition, how they cope with the slow evolution of markers typically used to characterize bacterial microbiota remains unclear. Here, we use simulations to test the statistical performances of two widely-used global-fit approaches (ParaFit and PACo) and two event-based approaches (ALE and HOME). We find that these approaches have different strengths and weaknesses depending on the amount of variation in the bacterial DNA sequences and are therefore complementary. In particular, we show that ALE performs better when there is a lot of variation in the bacterial DNA sequences, whereas HOME performs better when there is not. Global-fit approaches (ParaFit and PACo) have higher type I error rates (false positives) but have the advantage to be very fast to run. We apply these methods to the gut microbiota of primates and our results suggest that only a small fraction of their gut bacteria is vertically transmitted.

Dates et versions

hal-03873071 , version 1 (26-11-2022)

Identifiants

Citer

Benoît Perez-Lamarque, Hélène Morlon. Comparing different computational approaches for detecting long‐term vertical transmission in host‐associated microbiota. Molecular Ecology, 2022, ⟨10.1111/mec.16681⟩. ⟨hal-03873071⟩
19 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More