Evaluate bins with BUSCO in MOSHPIT tutorial 4.1.6 section

Technical Support
29

Hi @ferflei,
Thank you so much for sharing your data. I was able to figure out the issue! Thank you for your patience and help as we debug this.

So the TLDR is that when we are partitioning your samples, your completely numeric sample id's are getting filtered out because they don't match the string version that QIIME 2 is expecting and this is resulting in QIIME 2 thinking you have an empty manifest.

A simple workaround is to append a letter to all sample_ids and all directories from that data folder you shared. So 3019 will now be S3019 in all the MANIFEST file and the directory inside data/. (One thing do not chose E as your letter here because that's technically a valid number because of exponents. )

You will lose provenance this way but it will get you going immediately on your analysis.

Here is my code change that will fix this issue in an upcoming QIIME 2 release if you want to stay up-to-date: ids are always strings by cherman2 · Pull Request #398 · qiime2/q2-types · GitHub

Thank you again!

30

Dear @cherman2

Thanks a lot for your help to solve this issue, I did what you suggested and finally worked!

Amaizing! thanks again.

Fernando.

31

Hi All,

I wanted to circle back to this conversation to let y'all know that this bug has been fixed in the most recent release of Moshpit. Here is the Changelog for 2026.7: QIIME 2 2026.7 is now available!

Thank you again for catching this bug.

32

id test that sample id rename from the first reply before going through version updates - takes couple minutes and might save you a whole install cycle if it works