Sorry, just saw this. I know you already spoke to Steven about this, but I have no idea what you’re talking about with “CTOT” and “CTOB”! Steven’s been the one dealing with the bisulfite mapping. Thanks for the heads up, though! Does all this interaction with you on this project mean you’re going to come back to our lab?

Yeah, Steven did a de-novo assembly with the BS-seq data because reads weren’t mapping to the C.gigas genome when using BS-Map. I took the de-novo-assembled BS-seq data (see this notebook entry) and BLASTed against the C.gigas genome and the results were not good. So, this particular BLAST against the GenBank nt DB was to potentially see if our sequencing data was actually “our” data by attempting to ID which species the data matched with the most. So, this wasn’t a “real” attempt at data analysis, it was mostly a long shot at seeing if our data was mostly comprised of another species, thus potentially being the wrong data set sent to use from the sequencing facility.

I haven’t done a BLAST with just raw sequencing reads (i.e. not already de-novo assembled into contigs), so not sure how it would play out. However, BLAST has an option (can’t remember if it’s “on” by default or if you have to use a flag in the BLAST arguments when running the command to enable it) that detects short query reads and takes their length in to account when evaluating database matches. This, in theory, should help ensure that matches using short query sequences are accurate. However, it probably would be prudent to run the BLAST and evaluate the top 10 matches (alignments, e-vals, bit score, etc) to see if the results agree with your expectations.