Minimum Alignment Sequence Length To Assign Taxonomy

[Mykhaylo Usyk]

Hello,

I am running an ITS pipeline and have a question regarding how blast assigns taxonomy. What is the minimum number of nucleotides of the representative OTU sequence needs to align to a reference sequence before it is considered a hit in the blast database. I ask because after looking at one representative sequences, which was 367 bp, and manually aligning it to the blast hit I found that only 62 bp of the query were aligning. The percent identity was high at 94% with an e value of 1e-24, but the majority of the query is not used in the alignment. So at what length is the aligned fragment considered too short to produce a valid result?

Best,

Mike

[Colin Brislawn]

Hello Mike,

This is a great question because it forces us to evaluate exactly how the blast taxonomy assigner in qiime is working. To do this, I took a look at the source code:

https://github.com/biocore/qiime/blob/master/qiime/assign_taxonomy.py 

While I see blast hits be filtered by e-value and identify, I don't see mention of coverage. This could explain your 61 bp alignment.

https://github.com/biocore/qiime/blob/master/qiime/assign_taxonomy.py#L469-L471  

If you use the SortMeRNA program for taxonomy search, you can specify coverage with --sortmerna_coverage:

http://qiime.org/scripts/assign_taxonomy.html 

I hope this helps,

Colin