hyunminkang / cleancall Goto Github PK

Dear Hyun,

Thank you very much for developing the CleanCall tool.
May I ask for your help to understand the proper usage of the tool and to resolve the errors that I encounter.

1. There are mismatches of ref allele in the generated pileup and genome reference file that actually abort cctools verify :

FATAL ERROR -
ref G do not match with

I see that few percent of the positions indeed do have mismatches, for example c versus C, A versus AG and so on. This happens mostly at low coverage sites. The mismatches are unique for each sample, of course. How could I overcome this?

2. There is a general question regarding the best scenario to find possible cross-contamination in a large number of samples, N=1000-2000.
   Is it ok to build N pileups for each individual BAM file with cctools pileup and then supply them to cctools verify via ped file ?

$ cat *ped
#FAM_ID IND_ID  FAT_ID  MOT_ID  SEX     MPU
24-0992 24-0992 0       0       0       tmp.pileup.24-0992.txt.gz
24-1155 24-1155 0       0       0       tmp.pileup.24-1155.txt.gz
24-1000 24-1000 0       0       0       tmp.pileup.24-1000.txt.gz

Will it work to detect the contamination of a sample A by sample B ?

3. Is it true that the same VCF file with variant frequencies (e.g., from HapMap) is used for the pileup --loci, verify --vcf, and genotype --invcf ?
   This is how it is in the Quick start usage info. I am not sure I understand where does the input VCF for genotype correction come from.

Many thanks in advance,
Vasily.

stderr.txt
stdout.txt

I got the following message after running make. It seems that the hard coded path /home/hmkang/code/working/cleancall/ is the problem.

CDPATH="${ZSH_VERSION+.}:" && cd . && /bin/sh /home/hmkang/code/working/cleancall/missing --run aclocal-1.11 -I m4 /bin/sh: 0: Can't open /home/hmkang/code/working/cleancall/missing Makefile:270: recipe for target 'aclocal.m4' failed make: *** [aclocal.m4] Error 127

Please let me know if it is an easy fix.

Could you please post instructions running CleanCall when the source of contamination is known? I can't seem to find the anywhere.

Hi (Thank you for answering my other question).
cctools genotype writes strange symbols into the vcf file.
An example bam, index and output files can be found at https://cb.skoltech.ru/~enabieva/for_cleancall/ (the browser may complain about the certificate). An example of an offending line is on line 80 (chr20:310376). The reference is ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/technical/reference/GRCh38_reference_genome/GRCh38_full_analysis_set_plus_decoy_hla.fa
The vcf file is the HapMap3.3 from the GATK Resource Bundle (ftp://[email protected]/bundle/hg38/hapmap_3.3.hg38.vcf.gz)

Another issue I'm having is that if I don't restrict genotype to a particular chromosome, (chr20 in the example), it creates a Makefile target that's called just ".tbi" and then complains about it. I think it might have to do with prefixing the chromosome name with "chr".

Hi there, Thank you for creating cleancall.
I have some problems when analysis RNA-seq data:
The result file is as follows.

How can I get the contamination results of RNA-seq? I only want results that don't need correction.
or is there any problem with my running method?
'''
prod 11:48:21 R290-1: /mnt/GenePlus001/prod/maxx/software/contamination4/dbsnp
$ cat dbsnp_pileup.sh
/mnt/GenePlus001/prod/maxx/software/cleancall/bin/cctools pileup -loci /mnt/GenePlus001/prod/maxx/software/contamination3/dbsnp_138.b37.vcf.gz -index 180022992FR1 -out 180022992FRdbsnp_pileup -ref /products/repos/prod/Akso/BNC_v2/BNC/program/NoahCare/db/alignment/tgp_phase2_flat/hs37d5.fa --run 1 &
prod 11:49:17 R290-1: /mnt/GenePlus001/prod/maxx/software/contamination4/dbsnp
$ cat dbsnp_verify.sh
/mnt/GenePlus001/prod/maxx/software/cleancall/bin/cctools verify -index /mnt/GenePlus001/prod/maxx/software/contamination4/dbsnp/180022992FRdbsnp_pileup.ped -o 180022992FR_verify -vcf /mnt/GenePlus001/prod/maxx/software/contamination3/dbsnp_138.b37.vcf.gz --run 1
'''

Hi there, Thanks for creating cleancall.

I describe below some problems that I'm having with the tool:

I ran cctools pileupas described in the documentation: -loci VCF file and --index tabular_file_with_sampleID_bam_path.

I get the following error message:

Finished generating CAPT Makefile Running 10 parallel jobs of CAPT make -f t.pileup.Makefile -j 10 /bin/sh: 1: set: Illegal option -o pipefail make: *** [t.pileup.sample1.txt.gz.tbi] Error 2

I opened the makefile (*.pileup.Makefile) and I was able to copy and run the lines manually. With exception of (echo "#FAM_ID IND_ID FAT_ID MOT_ID SEX MPU"; cut -f 1 bam.man | perl -lane 'print join("\t",$$F[0],$$F[0],0,0,0,"output2.pileup.".$$F[0].".txt.gz")') > output2.pileup.ped

That i believe it is necessary to edit it (?!) to where it was $$F[0] -> $F[0]

Here is the command I ran: cctools pileup --loci VCF --index tabular_file_with_sampleID_bam_path --out output2.pileup --ref REF_genome --run 10

So there are two possible problems: 1) Makefile is not running correctly after the cctools pileup call and (maybe missing to install something (?!)) and 2) maybe the way that pileup.ped is generated may be problematic.

Thank you!!!

Hello,

The help screens of this tool and verifybamID are almost identical? are the stats underlying the programs the same?

Thanks

Matt

Will cleancall work with CRAM files?