Hi all,
What’s the cutoff of RTSI when using MelonnPan? In other words, what RTSI value means the sample is well represented by the training model?
Hi @nightkid - we do not have a recommended cutoff for the RTSI. As described in our manuscript, RTSI scores are computed by sequentially seeking the highest correlation coefficient with the top principal components (PCs) explaining a majority of the variation in the training microbiomes, which means it can be interpreted similarly to a correlation coefficient.