Skip to content

Latest commit

 

History

History
143 lines (114 loc) · 6.71 KB

CHANGES.md

File metadata and controls

143 lines (114 loc) · 6.71 KB
Raw

PAV changes

2.4.4

  • Corrected bug in VCF output for inversions and VCF output format for other variant types.

2.4.3

  • Added "no_link_qry" configuration parameter to force PAV to re-write query (assembly) FASTA files if this parameter is set to True. By default (no_link_qry = False), an acceptable input FASTA file is linked (single compressed FASTA as input).

2.4.2

  • Bad container build, re-released as 2.4.3

2.4.1.2

  • VCF writes failed variants with a FILTER
    • INFO/COMPOUND lists larger variants that caused smaller ones underneath it to be removed.
  • Added "ignore_env_file" configuration parameter to be used by the Docker entrypoint script to avoid disrputing the container environment if a "setenv.sh" file is present (container should use it's own environment).

2.4.1

  • Final VCFs write to vcf/merged
  • Avoid FASTA copy to temp (symlink)

2.4.0.2 (development version)

  • Filter variants inside inverted core (uniquely inverted loci, inner INV coordinates). Previous versions filtered all variants inside inversions (outer INV coordinates)
  • Skip input FASTA re-indexing if indexes already exist

2.4.0.1 (development version)

  • Arbitrary number of named haplotypes and new assembly table format (old format supported).
  • No support for "asm_pattern" in config (was removed before this version but was left in documentation).
  • Removed "chrom_cluster" from configuration options.
  • Split alignment trimming into distinct "tig" and "ref" phases (BED for each).
  • Keeping dropped variants and adding FILTER column with PASS and non-PASS values. Added COMPOUND field (and filter name called COMPOUND) to annotate variants that are dropped because they are inside another variant (i.e. small variant in a large DEL, COMPOUND field points to the accepted variant covering the dropped one). Makes it easier to recover variants if a large false-SV is removed.
  • Added ".fna" and ".fnq" to recognized file types.
  • Moving to Snakemake profile. Removed threading parameters from config (map_threads, inv_threads_lg, inv_threads) Configure in the profile instead of parameters.
  • Added joblib to container dependencies, removed Snakemake version restriction
  • VCF default rule updates

2.3.4

  • Adding "COMPOUND" column for filtered variants annotating which variants it was nested inside of (e.g. small variant inside a large deletion). Can be used to recover small variants if a larger variant is removed from the callset.
  • Fixed bug that dropped large insertions.
  • Fixed a bug generating tracks for references with numeric chromosomes.

2.3.3

  • Fixed minor alignment trimming bug uncovered by the large deletion fix in 2.3.2 causes large variant detection fail.
  • Fixed bug in coordinates causing inversion searches at the edge of chromosomes to fail with an invalid index (-1).

2.3.2

  • Fixed large deletion bug.
  • Removed "fake home" from container run script, causes problems with some systems. Some tools in the container will use real home (e.g. .cache)

2.3.1

  • Fixed typo in Singularity run script from home directory caching issues (thanks @aidenlx)
  • Fixed bug in variant filtering
  • Updated documentation for native installs related to deadlocks (OPENBLAS_NUM_THREADS)

2.3.0

  • Fixed home directory caching issues.

2.2.9

  • Fixed issues with home directory caching (related to bug fixes in 2.2.7)

2.2.8

  • Fixed rules for generating UCSC tracks for PAV calls and alignments.

2.2.7

  • Fixed bug creating the callable region BED if large SV BED files are empty.
  • Eliminating the need for --writable-tmpfs (Singularity), fails for some kernels

2.2.6

  • SNV ID format is now: CHROM-POS-END-REFALT
    • Eliminated dash between REF and ALT (single base each), consistent with SV-Pop library and makes IDs more paresable.
  • Updated SV-Pop submodule.

2.2.5

  • Added support for systems without /dev/shm
    • Set "inv_threads" and "inv_threads_lg" to "1" in config.json. Slower, but will not try to multiprocess.
  • Fixed bug in variant calling ("Passing a set as an indexer is not supported.", #34)

2.2.4

  • Updating for LRA index file name change from ".mmi" to ".mms".

2.2.3

  • Minor SV-Pop bug incorrectly de-duplicates variant IDs.

2.2.2

  • Minor SV-Pop bug incorrectly parsed chromosome names with "." while versioning variants.

2.2.1

  • Fixed missing library path in scripts/density.py
  • Quieted merging output
  • Minor documentation changes

2.2.0

  • Dockerized PAV
  • Added a small example (EXAMPLE.md)
  • Fixed bug writing VCFs with no variants of a certain type (most often caused by no inversions).
  • Fixed bugs when reference has numeric chromosomes (e.g. "1" instead of "chr1")

2.1.0

  • Added parameter "inv_inner" (boolean, default False) to allow small variants inside inversions
  • Added tunable parameter "minimap2_params" (thanks to Edwmard Rice).
  • If contig name is a number, convert it to str (thanks to Peng Jia).
  • Inversion caller bug fixes:
    • Incorrectly calling large inversions based on misaligned contigs.
    • Small inversions were being missed (scripts/density.py wasn't able to find kanapy, also fixed reason it was failing silently).
  • Batching haplotype merges (20 batches) instead of one per chrom (generates too many jobs on unplaced/unlocalized contigs)

2.0.1

  • If hap and parent are missing in asm_pattern, it's an unphased assembly (no longer an error).
  • Fixed INV signature bug when there are no indel clusters flagging INVs.
  • Reading small indels that cross Ns (particularly "NA") causes Pandas to read the table back with missing data instead of a "NA" deletion or insertion
  • Merging info changes (removed HAP_AC and HAP_AF, not meaningful).
  • Updated parameters for redundant callsets (multiple haplotypes in one assembly).
  • Added missing code for creating dotplots.
  • De-duplicating IDs before merging (for redundant callsets).
  • Fixed sys paths in scripts/reconstruct_sam.py.

2.0.0

  • Haplotype merging uses sequence content by default (not just size/position).
  • Fixed tuple assignment bug while finding input files (#12, thanks to Edward Rice).
  • Fixed SNV merge strategy typo (#14).

1.2.2

  • More granular merging control based on variant type.
  • VCF SVLEN header had the incorrect data type.

1.2.1

  • Changed parameter "merge_align" to "merge_match" to better match matching paradigm.

1.2.0

  • PAV merges using sequence context by default. Sequences are only merged if they are 80% identical by alignment parameters. Merging is done with SV-Pop parameters "match=0.8,2,-1,-4,-0.25,500000,9" (80% identical, match=2, mismatch=-1, gap-open=-4, gap-extend=-0.25, align up to 500,000 bp (Jaccard similarity if larger), Jaccard k-mer size 9). Parameter is tunable with option "merge_align", which may be "true" (default), "false" (no sequence matching), or a string that replaces everything after "match=".