search

cortes-ciriano-lab / savana

Somatic structural variant caller for long-read data
Apache License 2.0
43 stars 2 forks source link

working of savana cna calling #44

Closed sachingadakh closed 1 month ago

sachingadakh commented 1 month ago

Hello, Thank you for your suggestion on my previous issue #42. As per your suggestion in #42 , I ran the "savana cna" with the following command and received this error. Should I perform SV calling and phasing? However, I cannot because I do not have a normal sample, right? Besides, you had asked about the coverage and average read length in issue #42 . The coverage is around 1.23X by samtools depth and 1X by mosdepth. The average read size is 1.1 kb.

savana cna -t ../methyalation_data/dorado_meth_data/tumor.bam --outdir /disk2/ONT_files/ONT_methylation_analysis/CNV_data/savana/ --ref ../hg38.fa -w 1 Source: /home/rakieta/anaconda3/lib/python3.11/site-packages/savana/savana.py

... Allele counter will use threads = 24. (threads = 48 defined; threads = 24 available) ... Extracting phased heterozygous SNPs from None ... Traceback (most recent call last): File "/home/rakieta/anaconda3/bin/savana", line 10, in sys.exit(main()) ^^^^^^ File "/home/rakieta/anaconda3/lib/python3.11/site-packages/savana/savana.py", line 496, in main args.func(args) File "/home/rakieta/anaconda3/lib/python3.11/site-packages/savana/savana.py", line 193, in savana_cna allele_counts_bed_path = allele_counter.perform_allele_counting(outdir, args.sample, args.phased_vcf, args.tumour, args.allele_mapq, args.allele_min_reads, args.threads) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/rakieta/anaconda3/lib/python3.11/site-packages/savana/allele_counter.py", line 150, in perform_allele_counting het_snps = extract_hets(phased_vcf) ^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/rakieta/anaconda3/lib/python3.11/site-packages/savana/allele_counter.py", line 24, in extract_hets vcf_reader = cyvcf2.VCF(phased_vcf) ^^^^^^^^^^^^^^^^^^^^^^ File "cyvcf2/cyvcf2.pyx", line 278, in cyvcf2.cyvcf2.VCF.init File "cyvcf2/cyvcf2.pyx", line 205, in cyvcf2.cyvcf2.HTSFile._open_htsfile OSError: Cannot open '<class 'NoneType'>' for writing.

cmsauer commented 1 month ago

Sorry, I think my previous answer was a bit misleading. It seems that your data is inappropriate for this approach. For reliable SV and SCNA calling, you need whole genome (not adaptive sampling) sequencing data at ideally ~20-30X (or higher) with matched normal sequencing data available too. Unfortunately, you seem to have shallow sequencing data, and no matched normal. Hope this makes sense! Carolin