Our samples had been analysed by Mothur + Silva 119 reference DB around 2015. When analysing the same samples in 2023 with CLC genomics workbench and SILVA 119 DB, the results are strikingly different. The difference is obviously caused by the CLC and mothur pipelines, because the reference database is the same, but I am unable to understand the reason behind these differences.

Example of one difference: The prevalence of 'Streptococcus' genera across our samples was 100% with Mothur, but with CLC it has come down to ~40%.

Anyone face a similar challenge when updating methodology? And can help me make sense of these differences?

P.S. I had originally used a newer DB (Silva 138.1) for sample analysis and having found huge differences, I tested it with the older DB (Silva 119) to check the reproducibility of the results.

Thank you in advance