marbl / chm13-issues Goto Github PK

Hello,

I have issues with gff annotation using bcftools. At first, I used the following command
bcftools csq -f chm13_grch38_Y_added.fa -g chm13.draft_v1.1.gene_annotation.v4.gff3 ../longread/samtools/T15/T15_chm13_sorted.vcf -O z -o T15_chm13_sorted_annotated.vcf
then got the error message saying Could not parse the line, "Parent=transcript:" not present so I adjusted gff file using zcat chm13.draft_v1.1.gene_annotation.v4.gff3.gz \ | gawk 'BEGIN{IGNORECASE=0} $0 ~/^#/ || ($3 != "gene" && $3 != "transcript" ) { $0=gensub(/Parent=(.+?_T[^;]+)/, "Parent=transcript:\\1", "g", $0); print; next} $0 !~/^#/ { gsub(/ID=/, "ID="$3":"); $0=gensub(/Parent=(.+?_G[^;]+)/, "Parent=gene:\\1", "g", $0); print }' \ | bgzip -c > chm13.draft_v1.1.gene_annotation.v4.fix.gff3.gz to fix the error message.
But then I still get the phase error message
Error: GFF3 assumption failed for transcript CHM13_T0000003, CDS=111940: phase!=len%3 (phase=2, len=379)
and got the following message after --force running it.
Warning: GFF3 assumption failed for transcript CHM13_T0110841, CDS=23883573: phase!=len%3 (phase=0, len=607)
If bcftools seems to cause phase errors which tool do you recommend to use to annotate variant calls on CHM13?

Seems like others are experiencing the same issue here

marbl/CHM13#31

Cheers,
DK

We found excessive clipping in the Winnowmap alignment, particularly in the HSat2 and 3 arrays.
We are investigating this using alternative alignment methods and expect to have an update to the issues.bed track.

I had two SV Calling resultings from Pariament2 and sinffle software showing in blow:
Pariament2:

sinffle:

Frist,there are no base information instead of “DEL001207SUR” in the pariament2 resulting，how to transform it into base format just like sinffle resulting.
Second,how to merge two SV calling resulting according to your SVs format showing blow?

Which workflow / tools do you recommend to follow after aligning and filtering using Winnowmap? Should I follow standard minimap2 workflow after aligning? Thanks for your amazing works!

I generate an assembly with Peregrine2021 (0.4.1, main:eb7a2dc+) and find it may be useful for identifying some potential issues in the current T2T assembly although it may be equally likely to be issues in my assembly results.

• Method summary:
  • Generate independent assembly (not using T2T’s overlaps and graphs to avoid ascertainment biases)
  • Try to separate the assembly results to two haplotypes assembly-wise
  • Call variants with dipcall, generate a VCF file for the variants seen between the assembly and T2T chm13.draft_v1.0.fasta
  • Going through some dense variant call clusters and check if there are alignment anomaly manually
• For the pick 10 regions picked, four have nearby low coverage alignment signal from the bed files. Six do not. The alignment files in the IGV used are from markerandlength.idxsession (from Arang, the polishing team):
• I don’t have a systematic to generate a bed from from the VCF files that I have. However, if anyone is interesting, I can provide the VCF file.

Regions:
chr1:132,379,149-132,387,405
chr2:91,194,176-91,258,718
chr9:65,106,334-65,426,773
chr9:68,183,834-68,263,943
chr9:74,317,466-74,319,968
chr10:37,180,330-37,252,004
chr11:18,996,964-19,068,834
chr12:35,546,839-35,617,750
chr13:9,812,837-9,873,599
chr16:42,993,455-43,044,690

Thanks for this monumental work! Looks like it is v1 at http://genome.ucsc.edu/cgi-bin/hgTracks?db=hub_2395475_t2t-chm13-v1.0&lastVirtModeType=default&lastVirtModeExtraState=&virtModeType=default&virtMode=0&nonVirtPosition=&position=chr3%3A90755737%2D90789938&hgsid=1116564077_SGfUUmBADGIb8Vzqqk1VEwmZWELa

Also, wondering about the availability of GFF for the same?