Developed by Anand D. Patel in Vineet Bafna's lab at University of California, San Diego patel [dot] anandd [at] gmail [dot] com

InPhaDel takes as input, a scaffold of phased single nucleotide variants (vcf), whole genome sequencing data (sorted bam), proximity ligation data (sorted bam), and deletions (bed) and a classification model ('rf|svm|knn').

InPhaDel returns a prediction for the deletion being homozygous, scaffold A, scaffold B, or unlikely to be a deletion (bed). For details of the method, see Patel, Selvaraj, Bafna 2015.

Installing

Software Requirements

InPhaDel requires Python 2.7.x with packages numpy 1.9, pandas 1.13, scikit-learn 2.7.6, matplotlib [optional]. This tool may work on other version of pandas, scikit-learn, and numpy, but has not been tested.

The tool also requires samtools

Installing InPhaDel

InPhaDel can be run from source or installed.

Download the tar or zip ball and install using the setup.py script.

python setup.py install

To verify the installation/configuration is correct and run a small test case for InPhaDel (takes <5min) ::

python setup.py test 

If all is well, continue to usage.

Alternatively, to run from source directory use ::

python -m svphase.test.predict

Usage

InPhaDel requires inputs to follow a specific naming convention as described below,

a) Deletions must be in a simple bed file. Chromosome names must match reference, and below input bam filenames.

b) Reference in fasta file format (with fai index). Use samtools faidx to generate index.

c) Phased scaffolds in VCF file format, WGS, and HiC bam files split by chromosome. Some are required, and optional files will be generated.

• INPUT_DIR/wgs/all.stat - generated by samtools idxstat of composite wgs data [required]
• INPUT_DIR/hic/all.stat - generated by samtools idxstat of composite hic data [required]
• INPUT_DIR/wgs/CHROMOSOME.all.bam[.bai] - WGS data where .bai is generated by samtools index [required]
• INPUT_DIR/hic/CHROMOSOME.all.bam[.bai] - HiC data where .bai is generated by samtools index [required]
• INPUT_DIR/vcf/CHROMOSOME.vcf [optional if INPUT_DIR/wgs/CHROMOSOME.[AB].bam and INPUT_DIR/hic/CHROMOSOME.[AB].bam] have been generated.

INPUT_DIR is an argument to inphadel and CHROMOSOME is the name of a chromosome in the reference fasta file. There must be CHROMOSOME bam files for each CHROMOSOME in the bed file.

To see command line arguments, use ::

inphadel -h

or from source ::

python -m svphase.inphadel -h