Dear all,
I am posting this message because I need some help interpretation my results generated with Qiime2.
I attached to this message files that represent lefse and ancom generated to compare three different groups N/Y-/Y+.
Could someone give me some help understanding the generated files?
For example, could you give me a method to determine the W cutoff I have to apply ? In other files, hypothesis is considered as true with a 0 value, when in the attached files W=5 is considered as false.
Thanks in advance for your valuable help.
Best regards.
Maxime
Hi @Eolden,
I can’t speak with regards to the LEfSe analysis as I am not familiar with their output.
But the 2 ANCOM files you have attached:
ancom.tsv: The top 2 taxa with W scores of 100+ are the only 2 taxa identified to be different between your groups.
ancom_6_LOCAL_INFECTION.qzv: Only the Ruminococcaceae UCG-014 with a W score of 246 is considered different.
With q2-ancom, you shouldn’t really have to worry about setting your own W cut off score, this is done under the hood for each run. This isn’t like a p value whereby a fixed threshold can be set globally, so the same W thresshold can’t be used across different runs. You should stick with the predetermined significance threshold. The low W values (0 or 5 for example) appearing as significant is a known issue with the q2 version of ANCOM. Have a look at this thread for more details on that. You can also check out q2-ALDEx2 and q2-songbird for alternative differential abundance methods if ANCOM is giving you issues.