Illumina / canvas

Canvas - Copy number variant (CNV) calling from DNA sequencing data
Other
121 stars 20 forks source link

System.ApplicationException: ERROR: Expected to read 27029 bytes from the block header, but only read 20537 bytes. #128

Closed xxxue96 closed 5 years ago

xxxue96 commented 5 years ago

Hi,

Below is my command for canvas: mono $canvas/Canvas.exe Somatic-WGS --bam input.bam --b-allele-vcf $ref/canvas/dbsnp.vcf --exclude-non-het-b-allele-sites -n V300017817_L3-566 -o ./ -r $ref/canvas/kmer.fa -g $ref/canvas -f $ref/canvas/filter13.bed

My input.bam is mapping to GRCh38 with bwa mem and removing duplicates using picard. All reference files are downloaded from S3. Canvas version is v1.11.0. I tried canvas 1.40.0. However, it gave me an error "The specified vcf file () does not exist" which was mentioned in #127 . That's why I turn to use v1.11.0

Log and error files are attached: CanvasLog.txt CanvasError.txt CanvasSNV-V300017817_L3-566-chr11.stderr.txt

Does that mean there is something wrong with my bam file? If so, any suggestions for it? Thanks for help.

eroller commented 5 years ago

Can you use samtools view to verify the bam is properly formed? For Somatic-WGS you should use 1.40.0. The missing VCF is likely a problem with mismatching chromosome names or something wrong with the b-allele vcf. It is recommended that you use strelka to generate the b-allele vcf from the matched normal bam file.