v1.2

• Changed multi-processing procedure to read through the input file line-by-line and process chunks of input in parallel. This replaces the previous strategy of splitting the input file into separate files to process, which resulted in undesired duplications of large files and multiple reads of the input file.
• Improved logging to give more frequent updates as the program is running and better formatting of the updates.
• Removed the -i option that allowed for multiple input files to be specified in a single run. Equivalent behavior can be replicated by the user running Referee multiple times on multiple input files with job scheduling tools (such as SLURM or GNU parallel).
• Removed 100GB input file size limitation.

Referee v1.1.21

v1.1.21

• Added maxmimum input file size of 100GB to input -gl files to prevent users from duplicating large files and having intractable run times.
• Added the bam_split.sh and tab_split.sh helper scripts to aid users that need to split large BAM, pileup, or ANGSD files by scaffold.
• Added an error message when trying to read .gz compressed FASTA reference file.

v1.1

• Implemented the --haploid option to score genomes from haploid species. This option can only be used in conjunction with an input --pileup file as the likelihoods must be calculated internally. This restricts the possible genotype likelihoods to the four base states and only considers a single allele in the likelihood calculation. See: https://gwct.github.io/referee/calcs.html
• Fixed the order of operations for the probability calculations with mapping quality.
• Updated unichr to chr for Python 3.
• (v1.1.1) Fixed a bug where all scores output would be -2 after encountering a site with no reads mapped in a pileup file. Thanks to @reedacartwright.

To install

Referee does not need to be compiled. Simply download one of the source packages below, decompress, and run with python referee.py <options>. You may want to add the referee folder to you $PATH variable for ease of use.

Referee v1.1.1

v1.1.1

• Fixed a bug where all scores output would be -2 after encountering a site with no reads mapped in a pileup file. Thanks to @reedacartwright.

v1.1

• Implemented the --haploid option to score genomes from haploid species. This option can only be used in conjunction with an input --pileup file as the likelihoods must be calculated internally. This restricts the possible genotype likelihoods to the four base states and only considers a single allele in the likelihood calculation. See: https://gwct.github.io/referee/calcs.html
• Fixed the order of operations for the probability calculations with mapping quality.
• Updated unichr to chr for Python 3.

To install

Referee does not need to be compiled. Simply download one of the source packages below, decompress, and run with python referee.py <options>. You may want to add the referee folder to you $PATH variable for ease of use.

Referee version 1.1

• Implemented the --haploid option to score genomes from haploid species. This option can only be used in conjunction with an input --pileup file as the likelihoods must be calculated internally. This restricts the possible genotype likelihoods to the four base states and only considers a single allele in the likelihood calculation. See: https://gwct.github.io/referee/calcs.html
• Fixed the order of operations for the probability calculations with mapping quality.
• Updated unichr to chr for Python 3.

To install

Referee does not need to be compiled. Simply download one of the source packages below, decompress, and run with python referee.py <options>. You may want to add the referee folder to you $PATH variable for ease of use.

Referee version 1.0

Published implementation of Referee (https://doi.org/10.1093/gbe/evz088).

• Takes either pre-calculated genotype likelihoods or a pileup file for internal genotype likelihood calculation and uses these to calculate site-by-site quality scores for the input assembly.
• Incorporates mapping qualities into internal genotype likelihood calculations.
• Supports multi-processing.
• Outputs to tab delimited, fastq, and bed formats.

To install

Referee does not need to be compiled. Simply download one of the source packages below, decompress, and run with python referee.py. You may want to add the referee folder to you $PATH variable for ease of use.