GeMS 1.0 Overview: Genotype Model Selection (GeMS) is a single nucleotide polymorphism (SNP) detector that works with alignment files of high-throughput sequencing (HTS) data. GeMS calls SNPs based on a statistical model selection procedure and accounts for the errors that can occur during genomic sample preparation. Compile: Use a C++ compiler to compile gems.cpp. For example, users of GCC in a Unix-like computing environment can use the command: $ g++ gems.cpp -o gems to compile GeMS with the associated header files. Input: GeMS accepts input files in the SAMtools pileup format. To convert a SAM/BAM alignment file into the pileup format, users can use the SAMtools pileup procedure with option -s. Filter: Many HTS data alignment procedures record information about the short reads which are aligned to a reference genome. Short reads with undesirable characteristics can be pre-filtered before running GeMS. For an explanation why pre-filtering may be desirable and for a brief tutorial on how to pre-filter SAM/BAM alignment files using SAMtools view, please see the PDF document entitled "Pre-Filtering Alignment Files" available at https://sites.google.com/a/bioinformatics.ucr.edu/xinping-cui/ home/software/ngs-snp-calling. Usage: gems -i input.pileup -o gems_output.txt [OPTIONS] Options: -d 0/1 0 indicates diploid, 1 indicates haploid, default is 0 -s FLOAT maximum likelihood computing steps float value, smaller is slower yet more precise, default is 0.001 -m INT maximum number of alleles to be considered at each site, 0 indicates unbounded, default is 255 Output: The GeMS output consists of 6 columns: 1. Chromosome identifier (character string) 2. Reference site on chromosome (integer) 3. Reference allele (single character) 4. Consensus allele (single character) 5. Largest posterior probability of the genotype models (float) 6. p-value of Dixon's Q test (float) All sites which pass a SNP pre-filter are output. The sites that qualify as GeMS SNP calls are those in which the reference allele differs from the consensus allele and the Dixon Q test p-value is less than the user specified alpha value. Contact: Xinping Cui https://sites.google.com/a/bioinformatics.ucr.edu/xinping-cui/