PUNG further analyzes polymorphic killer cell immunoglobulin like receptors (KIR) short-read sequencing data results analyzed with PING, and discriminates potential new alleles sites from annotation errors to streamline new allele validation.
We developed an automated pipeline that mines large dataset outputs from PING to determine whether unresolved genotypes are a product of low-quality sequence data, misaligned reads from another KIR gene, or new alleles. The pipeline performs quality control of new variants, determines if the novel variants are specific to the gene being analyzed, and it aligns reads to a local database containing all KIR variants deposited on the ImmunoPolymorphism Database (IPD-KIR). Finally, PUNG verifies whether the resulting sequences have already been deposited in IPD-KIR or are suggested new alleles.
From our validation dataset, PUNG was able to completely resolve 35.6% of the unresolved genotypes in the dataset. Using Sanger sequencing, we confirmed 81.1% of PUNG suggestions, and 21 newly discovered KIR3DL2 alleles were described while minimizing computational requirements. Thus, we suggest that PING be used in conjunction with PUNG to enable state-of-the-art study of KIR polymorphism.
PUNG currently supports 9 of the 13 KIR loci (KIR3DL3, KIR2DS2, KIR2DL1, KIR2DP1, KIR3DP1, KIR2DL4, KIR2DS1, KIR2DS4, KIR3DL2).
Author: Luciana Vargas
Last updated October 6, 2022
Requirements: Linux
Download the PUNG program to your machine.
git clone https://github.com/lucivargas/PUNGIn the PUNG_master.R files, modify the input and output variables (lines 8 and 9), and source the script.
PUNG produces the following output files:
The finalAlleleCalls_PUNG.csv contains an updated genotype data frame.
The NewAllelesList.csv contains a data frame with the following columns: ind: Sample name locus: Locus name genotype: Genotype found in PING results copy_number: Copy number of the locus in that sample genotype_string: Final genotype after analysis with PUNG snp_list: Final list of suggested new alleles after analysis with PUNG, if any
The log file (.log) contains a summary of the analysis.
The snp_output folder provides updated tables of SNP calls for each locus after analysis with PUNG.