REQUIREMENTS: NAHR_finder must be run on compute nodes that are MPI-enabled and have a CUDA-capable GPU. Additionally, NAHR_finder utilizes the following packages. CUDA >= 5.5 BOOST C++ libraries >= 1.55 with additional packages: serialization, mpi, system, filesystem, statistical distributions GMP MPFR MPFRC++ GMPFRXX BAMtools MATLAB INSTALLATION: After installing the required packages, the "makefile" must be manually changed to link to these packages. Then, you need simply run: make USAGE: detect-NAHR [options] options: -d a database of segmental duplications (SDs) in the reference genome and pairwise alignments amongst homologous SDs. [the default database that comes with this package is for hg19 (human).] -R haploid reference genome fasta file prefix for each chromsome [default: hg19_chr] -b input BAM file of paired-end reads aligned to the reference genome -s path to a temporary directory for temporary files -o output directory -varpos path to a directory which contains a file for each chromosome, where each file contains all of the positions in that chromosmome that are on an SD and are a variational position. -l length of reference genome [default: 3190491286 for hg19] -e base error-rate for the paired-end reads. Context-specific error-rates adjustments will be made against this base error rate. [default: 0.01] -p a priori probability of no NAHR or gene conversion event occurring at a given locus [default: 0.999999] -compute_size comma-separated pair of integers. Connected components will be selected for computation if their compute size is within this range. [default: 0,7000] -alpha read-depth variance factor. (Let u_R = expected read-depth of a region R. Then the likelihood of observing n reads from region R is N(n ; u_R, sigma^2) where sigma^2 := u_R * (1+alpha) [default: 0.01] -samples number of samples from posterior distribution to draw to estimate posterior probabilities. -vs path to directory containing visitation schedules for connected components -gender_and_pop file containing tab-delimited information about the gender and population of the individual sequenced in the BAM file. -G name of individual. CONTACT: Please direct all questions and problems to: matthew_parks@alumni.brown.edu