specifying "GI mapping files" for MEGAN5 on command line interface

Question

[JV]

I like to use MEGAN5 for analyzing bins sorted from assembled metagenomes.

I since I have Blast results (against nr) in tabular format, I usually just use the "Import from BLAST"-option from the "File" Menu and specify a GI-mapping file to get obtain the taxonomy info for the blast results. (Similar I would use SEED- and KEGG- mapping files to obtain functional info).

Since I have a large number of bins that I want to analyze, it is getting rather tedious to click through the menus by hand.

Therefore I want to use the command-line interface of MEGAN5 to automize those tasks.

Loading the gi mapping file seems pretty straightforward:

load taxGIFile=/usr/local/share/db/MEGAN5_mapping/gi_taxid_prot.bin

Importing the Blast-File is also pretty straightforward:

import blastFile=blastp_nr.out.tab fastaFile=bin_79_allmarkers.faa meganFile=delmetest.rma

When I run the two above commands, first the mapping file is successfully imported, and then the BLAST-file is imported. However the taxon names are not being mapped, resulting in 100% unassigned reads.

Probably I have to use the mapping=<mapperConfiguration> argument when importing the blast file, but I have no idea what to give as actual argument. I can't find any info on how to set the "mapperConfiguration". I've tried simply passing mapping=YES as argument, but that didn't help.

Does anyone have experience on using MEGAN5 though the command-line interface?

[bedeabc]

Did you ever work this out? I have exactly the same problem. I see people using mapping='Taxonomy:BUILT_IN=true'; but it doesn't work for me.

Answer 1

[bedeabc]

Ok, I figured it out with thanks to http://seqanswers.com/forums/showthread.php?t=43535

The mapping command one needs to force MEGAN to use the previously imported GI mapping file is mapping='Taxonomy:BUILT_IN=true,Taxonomy:GI_MAP=true' in the import blastFile line

If only the MEGAN command line were properly documented and included even some basic examples, it would save everyone an awful lot of time.

[JV]

Thanks, I must have given up on this and forgotten about this thread, because i accidentally posted a similar question when I tried again some time later and ran into another problem: How to use scripts with MEGAN5 in commandLineMode. Turns out that after some time, I stumbled upon the same helpful SeqAnswers thread as you (but it would have been faster if I had noticed your reply first).