MEGAN5: parsing problem

Question

[GeJu]

I have an assembled transcriptome that I blasted against the nr_db from NCBI. I want to use MEGAN to get some taxonomic information on the reads. I imported my blast xml files into MEGAN5 and all seemed to work fine. When I click a node to inspect which hits are associated with a given taxon I encounter a problem. The majority of the hits are listed correctly with the query id header but some do not have a header and the hole sequence or parts of a sequence is displayed. This results in incomplete sequences when I want to extract subsets of reads.

I have no clue what is happening there and how to fix it. Anyone encountered a similar problem?

Could there be a problem with the blast xml input files? Or is it a parsing problem? Any idea how to fix this?

Any help is appreciated!

[ashishtx]

Is it showing any error message or just not displaying anything?

[GeJu]

No error message, I just noticed it because I extracted some sequences.

Answer 1

[GeJu]

I figure out what went wrong. It was a tiny formatting error in my blast input files that corrupted everything, thought blast had no problem with it. Fixed it and now everything is working fine!

[LauferVA]

nice work fixing it - please consider posting the incorrect input, the resulting error, the desired output, then the fix you used. it is very possible others will make a similar enough mistake in the future to make it worthwhile.

good job on figuring it out :-)

Answer 2

[5heikki]

Have you consider alternatives like blast2lca? xml output would be no good for it though..