Dataset for TS inference

Dataset for TS inference#

Diploid, phased VCF files with varying numbers of individuals. All are BGZIP-compressed and have a CSI index. A key to the populations is in the file popKey.

These are for TS inference with SINGER, threads, and tskinfer/tsdate. For each file, the REF allele can be assumed to be ancestral. The imputed files with Maj in their name have the REF allele set to the major allele. This is to assess the effect of wrongly specified ancestral alleles.

  • Complete VCF

    • tsm100M300I.vcf.gz (exported from msprime with ref allele = ancestral)

  • Array-density files

    • ts300I2k.vcf.gz

    • ts300I25k.vcf.gz

  • Imputed datasets for TS inference

    • panelLarge25kImputed.vcf.gz

    • panelLarge25kMajImputed.vcf.gz

    • panelLarge2kImputed.vcf.gz

    • panelLarge2kMajImputed.vcf.gz

    • panelNoBLarge25kImputed.vcf.gz

    • panelNoBLarge25kMajImputed.vcf.gz

    • panelNoBLarge2kImputed.vcf.gz

    • panelNoBLarge2kMajImputed.vcf.gz

    • panelSmall25kImputed.vcf.gz

    • panelSmall25kMajImputed.vcf.gz

    • panelSmall2kImputed.vcf.gz

    • panelSmall2kMajImputed.vcf.gz

    • panelNoBSmall25kImputed.vcf.gz

    • panelNoBSmall25kMajImputed.vcf.gz

    • panelNoBSmall2kImputed.vcf.gz

    • panelNoBSmall2kMajImputed.vcf.gz

Fix file names#

Files have been downloaded and extracted in data/toInfer/ folder: there’s a repetition of “vcf” in the file names. You can fix this by creating a simple script like this in the data/toInfer/ folder:

#!/bin/bash
# replace .vcf.vcf extensions with .vcf
# ex. panelLarge25kImputed.vcf.vcf.gz -> panelLarge25kImputed.vcf.gz
set -euo pipefail

for file in $(ls *.vcf.vcf.*); do
    newname=$(echo "$file" | sed 's/\.vcf\.vcf/\.vcf/')
    mv "$file" "$newname"
done

Make the script executable with chmod +x fix_filenames.sh and run it with ./fix_filenames.sh.

Create a TSV file with sample information#

Those files are a simulation of 1600 individuals from the original 2405 individuals in the msprime simulation. We need to create a TSV file with sample information using FID and IID columns: you can create it with an utility script in this folder:

python scripts/createFID-IID.py \
    --indiv-list data/toInfer/popKey \
    --directory data/toInfer

Create a fake FASTA file#

We need to create a fake FASTA in order to exploit bcftools reheader to add contig length information to the VCF files: this is required by create_tstree script installed in this project (and to be used by the cnr-ibba/nf-treeseq Nextflow pipeline). This can be done again with an utility script in this folder:

python scripts/fakeFastaFromVCF.py \
    --vcf data/toInfer/tsm100M300I.vcf.gz \
    --output data/toInfer/tsm100M300I.fa.gz

Call nextflow pipeline (tskit reference approach)#

You can create two nextflow configuration files for running the cnr-ibba/nf-treeseq pipeline: see config/samples_toInfer.csv and config/samples_toInfer-reference.json to see how to set it up. Then you can call the pipeline with:

nextflow run cnr-ibba/nf-treeseq -r v0.3.0 \
    -config config/custom.config -profile ibba,core -resume \
    -params-file config/samples_toInfer-reference.json

Call nextflow pipeline (threads approach)#

You need also two configuration files for running the cnr-ibba/nf-treeseq pipeline using threads approach, the config/samples_toInfer.csv file is the same as before, but you need to create a new JSON configuration file: see config/samples_toInfer-threads.json to see how to set it up. You can call the pipeline with:

nextflow run cnr-ibba/nf-treeseq -r v0.3.0 \
    -config config/custom.config -profile ibba,core -resume \
    -params-file config/samples_toInfer-threads.json

Call nextflow pipeline (threads with fit to data)#

You need to create another JSON configuration file for running the cnr-ibba/nf-treeseq pipeline with threads and fit to data option enabled: see config/samples_toInfer-threads-fit.json to see how to set it up. You can call the pipeline with:

nextflow run cnr-ibba/nf-treeseq -r v0.3.0 \
    -config config/custom.config -profile ibba,core -resume \
    -params-file config/samples_toInfer-threads-fit.json

Warning

The threads_fit_to_data option requires more resources (RAM and time) than the other pipelines, so make sure you have enough resources allocated.

Error

It was not possible to run the fit to data option on the smallest datasets (ts300I2k) using the current parameter set.

Contents