boulund/merge_taxonomic_classifications

Merge Kaiju, Kraken, and CLARK-S classifications of metagenomic reads

Merge taxonomic classifications

A small utility tool for merging Kaiju, Kraken, and CLARK-S classifications of metagenomic reads. Merges classifications in a user-specified order, overwriting assignments of classified reads made by methods earlier in the merge order.

merge_taxonomic_classifications's Issues

Summary unclassified read count incorrect when merging with empty file

If merging with an empty file, e.g. merging an empty Kraken output file with a Kaiju output file with merge order kraken,kaiju, the summary unclassified read count is incorrect.

The reason for this is that when reading the kraken file, no classified or unclassified reads are added to the SQLite3 database. Thus, when the kaiju file is read and "merged" with the kraken results, only the classified reads from kaiju are merged. It is assumed by the merge order that unclassified reads from kaiju are unclassified also in kraken, which is incorrect in this special case. This isn't really a problem, but still annoys me a bit :)

One of the earlier implementations did not have this issue, as it stored a complete SQLite3 table for each input file, and THEN merged everything, wasting a lot of time and space.