Analysis showing unexpected species

[hawaii454-0]

Hello,
We have made TruSeq RNA and DNA libraries for projects sequencing different species and after the bioinformatics QC, the analysis is showing up sequence from species that really shouldn't be there like zebra fish, salmon, and human.

As this has happened over different projects over the past eight months, we don't understand why these same specific species have been showing up across libraries.

Is there a different type of analysis that can be done as part of the QC pipeline to avoid well-sequenced reference genomes like that of the zebra fish? Is there something in these specific genomes that is similar and will interfere with preps/analyses/etc? All of the projects this has occurred in were de novo. Would this affect non-de novo projects in a similar way? Does anyone have any experience of this?

Thanks,
Anna.

[pmiguel]

Hi Anna,
We see effects similar to the one you are describing -- much as you write. Some genomes seem to be much more exhaustively annotated and so you end up getting a large number of hits to them, even though a more closely related species genome might be available.

As to the larger question -- how to set up the pipeline to prevent these "annotation hog" genomes from swamping out other, more appropriate, annotations -- no idea. However, if you want to "prove" that the library you did not accidentally sequence the wrong library, you could use matching to a good rRNA database as a QC step. Alas, this is not as trivial as it sounds because the rRNA databases commonly available are not really designed with this purpose in mind. So they often are missing key ribosomal RNAs, which lead to exactly the same problem -- you end up getting hits to more distant species reported because the more closely related species element is missing from your database.

--
Phillip