yangjl / pseudoref

Make Pseudo-Reference Genome from VCF/BCF

pseudoRef

It is an R package to make a pseudo reference genome subsituted with the SNP variants of given samples.

Installation (< 10 mins)

Install devtools first, and then use devtools to install pseudoRef from github.

devtools::install_github("yangjl/pseudoRef")

It requires two other papckages:

library("data.table")
library("Biostrings")

Usage

You can find help documentation by simply typing ?pseudoRef, which is the major function have been implemented in this package.

Usage:

     pseudoRef(fa, snpdt, sidx = 5:ncol(snpdt), arules = NULL, outdir)

Arguments:

      fa: Path for the reference fasta file. [string or
          DNAStringSet/DNAString object]

   snpdt: A data.table object with heterozygote SNPs coded with IUPAC
          ambiguity codes. [data.table, 4 required columns: chr, pos,
          ref, alt, (sample1, ..., sampleN)]

    sidx: A vector to indicate the sample columns. [vector,
          default=5:ncol(snpdt)].

  arules: Additional nucleotide substitution rules defined by users.
          [data.frame, 2 required columns: from, to, default=NULL] For
          example, arules <- data.frame(from=c("M", "Y", "R", "K"),
          to=c("C", "C", "G", "T")).

  outdir: Output directory. Sample specific sub-folders will be
          created. [string]

Note

Before running the package, we should use BCFtools to convert VCF/BCF file into IUPAC coded tab seperated file:

# bcftools view JRI20_filtered_snps_annot.bcf.gz -m2 -M2 -v snps -Oz -o JRI20_bi_snps_annot.vcf.gz
# bcftools query -f '%CHROM\t%POS\t%REF\t%ALT[\t%IUPACGT]\n' JRI20_bi_snps_annot.vcf.gz > JRI20_bi_snps_annot.txt
# bcftools query -f 'chr\tpos\tref\talt[\t%SAMPLE]\n' JRI20_bi_snps_annot.vcf.gz > JRI20_bi_snps_annot.header

License

This package is free and open source software, licensed under GPL.

pseudoref's Issues

object 'SAMPLE' not found

I have been trying to use it, but I keep running into errors, and I would really appreciate your help.

The error I keep running into is this:

pseudoRef(fa=fa, snpdt= snpdt, sidx=sidx, arules=NULL, outdir="/work/alj25/MED84_pseudoref/")
Error in [.data.frame(x, i) : object 'SAMPLE' not found
I converted my fasta reference to a DNAstring object (using readDNAStringSet), my snpdt is a data.table from my bcftools output, as you describe. I think the problem may lie there, because 1. I had to add names to the columns after importing (I never used the .header file you specified. How do I incorporate that?) 2. I only have 1 sample in the file, not multiple, so I don't think I'm actually able to provide a vector (sidx=5; and I have tried 5:5, 5 as an int and num).

Your script looks like it replaces the sample column name with "SAMPLE" in a loop, so I'm not sure why this isn't working, but please let me know if you do. I have also tried naming that column "SAMPLE" from the start, but I keep getting this error.

BCFtools filteration criterion (biallelic sites)

Hi,

Thanks for the 'pseudoRef' commitment.

I noticed that you kept only biallelic sites by using '-m2 -M2' parameters in bcftools.
bcftools view JRI20_filtered_snps_annot.bcf.gz -m2 -M2 -v snps -Oz -o JRI20_bi_snps_annot.vcf.gz

There are other multiallelic sites in a large population, why do you discard those sites?