I am working with a group who would like some guidance on how to create custom mapping files based on an alternative database (in this case, UniRef90). They are working on a very large sample set so are trying to utilize UniRef to help speed things along (similar in respect to the last step used in the HUMAnN2 workflow).
Thanks,
chris
A mapping file can be supplied in tab-separated format, so just a list of pairs
ref-seq-id tab taxon-id
for taxonomic binning
and e.g.
ref-seq-id tab InterPro-id
for InterPro2GO functional mapping.
MEGAN will look for ref|XXX in a match header line to find the ref-seq id and will then lookup XXX in the mapping that you provide to determine the taxon or functional class associated with the match.
If the accessions follow some tag other than ref| then there is a way that MEGAN can be told that.