Fingerprint toolset

The fingerprint toolset provides functionality for in-silico genotyping. The toolset can be used to confirm expected genotype matches (e.g. tumor & normal, WGBS & 450k array), or to detect unexpected matches (e.g. different DNA sources, sample swaps).

This code is related to Volker Hovestadt's methylCtools.

Usage

Software requirements

This is legacy code. The software versions that probably work with this code are these

• Python 2.7.9
  • pysam 0.7.7
• R 3.4.0
  • minfi 1.24.0
  • getopt 1.20.0

Generate fingerprint files (.fp)

• From BAM alignment files (python script):

  python fingerprint/bsnp.py \
     fingerprint/snp141Common_RS-CG.n150.vh20151103.bed \
     sample.bam \
     > sample.bam.fp

• From 450k data files (R script, RData input file must contain a minfi RGset object of a single sample):

  Rscript fingerprint/asnp.R \
     fingerprint/snp141Common_RS-CG.n150.vh20151103.bed \
     sample.RData \
     sample sample.fp

Analyze fingerprint files to identify genotype matches

• R script, input arguments can be individual files and directories, separated by commata:

  Rscript fingerprint/bsnp_analyze.R \
     -i sampleset/,some1.fp,some2.fp \
     -o sampleset

More detailed information is provided with the help messages and headers of the tools. It is recommended to understand and adapt the settings used in bsnp_analyze.R. often it also helps to look at the pairwise correlation table directly (e.g. in Excel using conditional color formatting).

The bsnp_analyze.R tool follows a very simple statistical approach (i.e. just calculating the Pearson's correlation), but provides reasonable results. Because of that it requires SNPs that are highly variable within the population (avHet >0.2). Anyone is invited to come up with a more sophisticated approach.

Included SNP files

GRCh37/hg37/hg19

• snp138Common.n1000.vh20140318.bed: A set of 1000 SNPs (avHet >0.25), selected for coverage in RNA-, Exome- and ChIP-seq datasets, including 450k RS probes. This file is recommended for comparing sequencing data.
• snp141Common_RS-CG.n436.vh20151103.bed: The original set of 436 SNPs (major allele freq <0.95) that can be assessed on the 450k arrays (RS probes and informative CG type I probes).
• snp141Common_RS-CG.n385.vh20151104.bed: A subset of the previous (consistency with WGS data, i.e. some probes with very high/low intensities were removed).
• snp141Common_RS-CG.n192.vh20151103.bed: A subset of the previous (avHet >0.2). This file is recommended for comparing 450k data, and for comparing 450k to sequencing data.
• snp141Common_RS-CG.n50.vh20151104.bed: A subset of the previous (filtered for coverage in Exome- and RNA-seq data).

GRCh38/hg38

• snp138Common.n1000_hg38.vh20140318.bed: set of 1000 SNPs (avHet >0.25), selected for coverage in RNA-, Exome- and ChIP-seq datasets, including 450k RS probes. this file is recommended for comparing sequencing data. This is basically the same as snp138Common.n1000.vh20140318.bed except for (see here for a detailed comparison):
  • rs1610774 & rs1698682 are switched and containing fragment is inverted in hg38 compared to hg19
  • rs61739740 has merged into rs9780
  • rs13369115 has merged into rs10155413

Author

• Volker Hovestadt

This code is related to methylCtools.

Changelog

• 1.1 [3. November 2016]

  • Added GRCh38/hg38 files

• 1.1 [4. November 2015]

  • New Git repository for old code.
  • Volker Hovestadt, German Cancer Research Center, 2015, v1.1