Skip to content

Latest commit

 

History

History
79 lines (69 loc) · 5.49 KB

CHANGES.md

File metadata and controls

79 lines (69 loc) · 5.49 KB
Raw

v1.0.0

  • first public release

v1.1.0

  • added --simple option to report genotype result for each repeat locus on a single line as an alternative output format
  • fixed setup.py and requirements.txt to make sure pip install using Github URL works

v1.1.1

  • bugfix in tre.py

v1.2.0

  • bugfix: forgot to pass reads_fasta to extract_missed_clipped()
  • bugfix: min_support checking should use "greater than or equal to"
  • removed dependency on intspan
  • reduced min_len from 20 to 15 for comparing motifs using blastn
  • bugfix: keep track of genomic coordinates of all alignments to detect reads that truly span long reference repeats
  • assigned min_span explicitly in conditional instead of relying on initialized value - min_span does not change in some cases when runing genome scan for some unknown reason
  • added check_same_pats() in is_same_repeat() to check if one motif is subsequence of another in blastn matches
  • test data added
  • output BED file by default in addition to TSV (user provide output prefix instead of name). BED output simplied without details such as read names.
  • added --regions to specify (in BED format) specific regions for scanning
  • added --min_cluster_size (default=2) to separate from --min_support so that alleles with fewer read support can be segregated
  • remove optional usage of dbscan for clustering
  • improve boundary definition of long(kb range) repeat loci in genome scan
  • added --tmpdir to allow user to specify TEMP location for generating tmp files
  • allowed --min_cluster_size to be bigger than --min_suport for genotyping mode only

v1.3.0

  • strand column added to .tsv output to indicated strand of support read, read_start is recalculated based on read instead of alignment coordinate for negative strand sequences
  • blastn jobs deployed per locus for compareing target and detected motifs combined to one job per process
  • minimum coverarge, percent identity, and e-value added to blastn jobs when launched
  • when searching for repeat in insertion site, target_flank increased from 2000 to 3000
  • bedtools merge parameter d for merging detected insertion loci increased from 50 to 100
  • reads identified as clipped at repeat locus will not be examined again for same locus as if it is a full alignment
  • complete removal of temporary bed files unless --debug specified
  • straglr_compare.py added for comparing straglr results from test against control samples
  • Python3.8 now needed becasuse Scipy1.8 is needed for t-test in straglr_compare.py
  • new parameter trf_args to allow user specify his own preferred TRF settings
  • changed seeding method of Gaussian Mixture Model to "k-means++" to reduce CPU overhead
  • --working_directory added to allow user specify working directory, otherwise current directory will be used

v1.4.0

  • locus, coverage, actual_repeat, read_status columns added, repeat_unit renamed as target_repeat in .tsv output
  • report failed reads in .tsv output with reason given in read_status column
  • added reporting of unpaired clipped alignments in genotyping results
  • allow unpaired clipped alignments to initiate TRE search in genome scan
  • added --include_alt_chroms to include ATL chromosomes in genome scan; added code to ensure chromosome boundaries are not violated
  • get rid of --working_directory, all temporary files will be found in $TMPDIR (or location specified by --tmpdir) if --debug is turned on

v1.4.1

  • detection and reporting of partial repeats in genotyping now an option (--include_partials)
  • bugfix: forgot to add read status to output vector from regex (not TRF) examination of repeats
  • bugfix: assign partials to allele if size is between maximum and minimum read sizes, do not assign if size cannot be placed
  • changed closeness_to_end from 10 to 1kb for the "unclipped" end of clipped alignments
  • always go ahead to genotyping phase even if repeat in insertion cannot be matched against locus in reference genome (behaviour before 1.4.0)
  • skip read if it has multiple clipped alignments to same end
  • fixed test.bam (duplicate records for some reads) and updated test results

v1.5.0

  • output VCF file
  • added --sex parameter: when m is specified, clustering of alleles is limited to one cluster for chrX and genotype is automatically homozygous in VCF for chrX loci
  • support CRAM file

v1.5.1

  • option to put - to NOT specify intended motif (4th column in BED file) in genotype mode, detected motif will not be screened for agreement
  • --symbolic parameter to report ALT alleles in symbolic form - only enforced when motif of alternative alleles is the same as reference
  • CLUSTERING_FAILED added as (default) reason for failure to detect/report genotype
  • ALT assignment conditions changed to using interquartile range instead of all support read sizes and at least one copy number difference from REF
  • bugfix: alternative motifs checked for equivalence for proper count reports in VCF output
  • bugfix: size and copy range of alleles only include "full" supporting reads even if partials were included, but "partial" reads will be reported in support counts
  • bugfix: allele motif equivalence checked to perform proper tallying for determining consensus motif in variant
  • used read name as tiebreaker for choosing support read sequence as ALT
  • add extensions (.fa, .blastn, .bed) to temporary files to help debugging
  • MOTIF renamed to RU, COPIES to REF in VCF fields

v1.5.2

  • post-processing of GMM clustering results modified
  • VCFv4.5 on tandem repeats (<CNV:TR>) followed