bcftools isec not producing a sites.txt file

[stevekm]

(posted here https://www.biostars.org/p/9464238/)

I am running a bcftools command like this;

bcftools isec -p /data --targets-file targets.bed Sample1.vcf.gz Sample2.vcf.gz

And I am getting output files like this;

0000.vcf
0001.vcf
0002.vcf
0003.vcf
README.txt

There is no sites.txt file output.

Normally when I run other sets of samples in bcftools isec, I also get a file sites.txt which has just the genomic regions used for intersection, and which sample each variant was present in, like this;

$ head input/sites.txt
1       45799087        G       A       111
1       115256527       C       T       110
1       115256527       CTTG    TTTT    001
1       115256530       G       T       110
1       193099343       G       T       010
2       26101038        C       A       100

Since I am not getting this file output, its making it difficult to parse with my existing workflow. Instead, the README.txt file has a description this like;

This file was produced by vcfisec.
The command line was:   bcftools isec -p /data --targets-file targets.bed Sample1.vcf.gz Sample2.vcf.gz

Using the following file names:
/data/0000.vcf     for records private to  Sample1.vcf.gz
/data/0001.vcf     for records private to  Sample2.vcf.gz
/data/0002.vcf     for records from Sample1.vcf.gz shared by both       Sample1.vcf.gz Sample2.vcf.gz
/data/0003.vcf     for records from Sample2.vcf.gz shared by both    Sample1.vcf.gz Sample2.vcf.gz

Any idea why I am not getting the sites.txt file?

Using

$ bcftools --version
bcftools 1.9
Using htslib 1.9

[stevekm]

looking at the source code, it seems like sites.txt is only created when >2 input files are used;

bcftools/vcfisec.c

Line 401 in 06e3645

args->fh_sites = open_file(NULL, "w", "%s/sites.txt", args->prefix);

Is this a bug? It seems like sites.txt should be created no matter how many input files are used.

[pd3]

It is not exactly a bug, just an omission. But you are right that there is no reason not to create the file also for the Venn set type of output. This is now fixed. Thank you for reporting the issue.

By the way, you are way behind the latest version of bcftools, we are now at 1.12.

[stevekm]

thanks, is this change included in the latest release?

[stevekm]

on an unrelated note, it looks like the reason I was using bcftools version 1.9 is because version 1.12 appears to be incompatible with the other tools I need as well;

$ conda install -y \
bioconda::bcftools \
bioconda::vcf2maf \
bioconda::bedops

...

The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    bcftools-1.9               |       ha228f0b_4         807 KB  bioconda
    bedops-2.4.39              |       hc9558a2_0         9.1 MB  bioconda
    bzip2-1.0.8                |       h7b6447c_0          78 KB
    ca-certificates-2021.4.13  |       h06a4308_1         114 KB
    certifi-2020.12.5          |   py37h06a4308_0         141 KB
    conda-4.10.1               |   py37h06a4308_1         2.9 MB
    curl-7.71.1                |       hbc83047_1         140 KB
    htslib-1.9                 |       ha228f0b_7         1.5 MB  bioconda
    krb5-1.18.2                |       h173b8e3_0         1.3 MB
    libcurl-7.71.1             |       h20c2e04_1         305 KB
    libdeflate-1.0             |       h14c3975_1          43 KB  bioconda
    libedit-3.1.20210216       |       h27cfd23_1         167 KB
    libgcc-7.2.0               |       h69d50b8_2         269 KB
    libssh2-1.9.0              |       h1ba5d50_1         269 KB
    ncurses-6.2                |       he6710b0_1         817 KB
    openssl-1.1.1k             |       h27cfd23_0         2.5 MB
    perl-5.26.2                |       h14c3975_0        10.5 MB
    samtools-1.7               |                1         1.0 MB  bioconda
    vcf2maf-1.6.21             |       hdfd78af_0          37 KB  bioconda
    ------------------------------------------------------------

^ conda chooses bcftools 1.9

Trying to install bcftools 1.12;

$ conda install -y \
bioconda::bcftools==1.12 \
bioconda::vcf2maf \
bioconda::bedops

...

Found conflicts! Looking for incompatible packages.                                                                                          failed

UnsatisfiableError: The following specifications were found to be incompatible with each other:

# a whole bunch of packages

[pd3]

This will be included in the next release, which is imminent.