Hi there!!! @franzosa @fbeghini @lauren.j.mciver @Kelsey_Thompson -
I am using kneaddata only for decontamination by bypassing
trf steps. I have run my control and test datasets in two different computers where the two kneaddata versions are different, i.e. 0.7.4 and 0.7.10. The outputs were analysed by HUMAnN 3.0.
Now, considering the two different versions of kneaddata, can I compare the control and test HUMAnN output? Or, I should re-run the whole process again with exactly same versions of kneaddata and again run with HUMAnN?
Hello, I just checked the version history log for Kneaddata and I don’t see any differences that would affect just the decontamination (bowtie2) step between those two versions. If possible I would always recommend running control and test datasets with the same version of all software in the same compute environment if possible just to make sure that all settings and dependencies are consistent. However, sometimes I understand that might not be possible so for your case if everything is the same but those two versions differ I think likely you should be okay. If possible, you could rerun just one data set on either platform just to double check all is as expected.