k-mer count > reference about bcalm HOT 6 CLOSED

Hi Simone,

1. It makes sense that BCALM outputs way more unitigs than reference contigs, because each unitig is broken as soon as there is a repeated k-mer in the reference with a different base after and/or before.

2. Regarding the small difference in kmer counts, this is odd. Could you please run a third kmer counter, e.g. KMC3 https://github.com/refresh-bio/KMC to see if it agrees with one method or the other?

Rayan

from bcalm.

Hi Rayan,

1. Yes absolutely, but I found weird that most unitigs are of length ~k, and that for k=25 I got almost 3 times the size of the reference. Still, this is of course possible.
2. I will run it and post the results. UPDATE: The k-mer count given by KMC agrees with jellyfish's

Thank you for your quick answer!
Simone

from bcalm.

Hi Simone,

Regarding 2. wow yes, indeed. Thanks for reporting it, I am investigating this discrepancy in DSK counts.

from bcalm.

No problem, let me know if you need more infos to reproduce the error

from bcalm.

Hi Simone,
Apologies for the delay in addressing this bug. It has stumped me for a while because that test case was quite large. Finally I found the issue: the reference FASTA file that you provided contains IUPAC nucleotides:
...caaTGGGCTTTCCATATATGGCCTTAATTATGTTGAGGTAGTTTCCTTCTGTTCCTAGTTTGGTGRATGTTTTTTATCAT.....

and BCALM doesn't handle them so well. It replaced the 'R' with a 'C', and shouldn't have done so.

from bcalm.

thanks again for the detailed bug report!

from bcalm.