Performance comparisons of single taxon placement and multiple taxa insertion.

For single taxon placement, the ranking is considering both placement accuracy and RF distance for the performances of TIPars, UShER and EPA-ng as presented in Fig 1. For example, in 16S dataset, though TIPars and EPA-ng achieved the same placement accuracy (81.3%), EPA-ng had a smaller mean RF distance (0.16) than TIPars (0.27) and would be ranked as better. For multiple taxa insertion, the ranking is considering both tree log-likelihood and RF distance for the performances of TIPars, UShER and IQTREE2 as presented in Fig 3. A caveat is that we have not compared IQTREE2 with RF distance because it may alter the input tree topology after inserting sequences, which would lead to a bias to its high tree log-likelihood. ’A > B’ presents A is better than B. (XLSX)