MEGAN5 help - no assigned reads

Question

[jkrauel]

I'm trying to use MEGAN5 to get order-level taxonomy assignments for CO1 reads. I have a set of OTUs in a fasta file. I downloaded all the CO1 records from NCBI and created a blast database, then did a local blastn against that using outformat 7. I am able to successfully import all this into MEGAN but it has zero assigned reads. I have played with the LCA parameters, and I'm currently using min score=1, max expected=100, min support=1. These seem to have little effect on the outcome. The blast output file records look like this:

OTU_1 gi|414148152|gb|JX897403.1|90.05 211   21    0     1     211   9219 1e-76    286

So I think the record identifiers are present. Below is the output from MEGAN. I read through the MEGAN manual and I can't really see anything else obvious to try. Can anyone give me any hints?

Executing: import blastFile='resultsall.txt' fastaFile='CO1ALL061015.fasta' meganFile='resultsall.rma' maxMatches=100 minScore=1.0 maxExpected=100.0 topPercent=10.0 minSupport=1 minComplexity=0.0 useMinimalCoverageHeuristic=false useSeed=false useCOG=false useKegg=false paired=false useIdentityFilter=false textStoragePolicy=Embed blastFormat=BlastTAB mapping='Taxonomy:BUILT_IN=true';
Importing data: 1 reads file(s), 1 blast file(s)
Input format: BlastTAB
TextStoragePolicy: Embed matches and reads in MEGAN file
Mapping all reads in memory
Processing CO1ALL061015.fasta: 33638
Processing resultsall.txt
Executing: show window=message;
Total reads:   656
Total no-hits: 0
Total matches: 332456
Matches discarded: 266207
Parsing required 12 seconds
Number of reads: 656
Low complexity:  0
With valid hits: 656
Number of taxa identified: 1
Data processor required: 1 secs
Total reads:        656
Assigned reads:     0
Unassigned reads:   656
Reads with no hits: 0
Reads low comp.:    0
Induce Taxonomy tree, keeping 2 of 1266115 nodes
Induced taxonomy tree has 2 nodes

[Istvan Albert]

ensure that the taxonomy mapping is also loaded

[jkrauel]

Thanks, Istvan, but I'm not sure what this means and I can't figure it out from the manual. I downloaded the gi_taxid-March2015X.bin file (3GB!) and it is in the same directory as the application. I am using the blast output format that should include adequate information for looking up the data. What else do I need to do? If this is documented somewhere please just point me to it.

Jennifer

Answer 1

[JV]

Your blast-results seem to be in tabular format and include the Gene-Identifiers instead of the descriptions. However that should be no problem (I only use tabular blast-outputs also). Did you load the gi-mapping file to MEGAN?

When importing your BLAST-results into MEGAN you need to click on the "Taxonomy"-Tab adn then tick the box "Use GI MAP". Then you have to click on the Folder-symbol next to "Load GI Mapping File" and point it to the location of your mapping file.

Please note that there are different mapping files for protein and nucleotide results, and that there are further different mapping files for mapping GIs to funCtions (e.g. Kegg), so please make sure that you have the right mapping file.

If you choose the (less human-readable and much larger) standard- or XML-outputs from blast, then MEGAN should be able to parse the taxon-information directly from the descriptions without using GI-mapping files (in that case just tick "Parse Taxon Names" in the "Taxonomy"-Tab when importing BLAST-results).

[jkrauel]

Thank you so much, JV. I had not realized I needed to load the GI map via the Taxonomy tab; I thought putting it in the application folder would be enough. Now it works just as I expected.