I have a couple of datasets like the below where it looks like there is an initial, po

Thanks, <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-

Thanks <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-u

premature kmer cutoff about hifiasm HOT 4 CLOSED

chhylp123 commented on May 27, 2024

premature kmer cutoff

from hifiasm.

Comments (4)

lh3 commented on May 27, 2024 1

Thanks, @mcshane. I have merged the fix to the master branch.

from hifiasm.

lh3 commented on May 27, 2024

Shane, you are definitely right that hifiasm misinterpreted this k-mer histogram. I have implemented a fix at the "fix-49" branch. Could you have a look to see if it really fixes the issue? Thanks!

PS: the fix skips the lower end of the k-mer spectrum when hifiasm attempts to look for the homozygous peak. The default is set to 5.

from hifiasm.

mcshane commented on May 27, 2024

Thanks @lh3 . That seems to have done the trick.

[M::ha_analyze_count] left: none
[M::ha_analyze_count] right: count[43] = 109907
[M::ha_pt_gen] peak_hom: 43; peak_het: 21
[M::ha_pt_gen::526.601*6.66] ==> indexed 202457353 positions
[M::ha_assemble::4091.051*[email protected]] ==> corrected reads for round 1
[M::ha_assemble] # bases: 8221544631; # corrected bases: 18234233; # recorrected bases: 51145
[M::ha_assemble] size of buffer: 3.867GB
[M::ha_pt_gen::4154.483*14.76] ==> counted 7997683 distinct minimizer k-mers
[M::ha_pt_gen] count[4095] = 0 (for sanity check)
[M::ha_analyze_count] lowest: count[10] = 77326
[M::ha_analyze_count] highest: count[21] = 206757

from hifiasm.

mcshane commented on May 27, 2024

Thanks Heng. It also worked on the other few datasets with the same issue.

from hifiasm.

Recommend Projects