Skip to content

Latest commit

 

History

History
118 lines (94 loc) · 4.88 KB

CHANGES.md

File metadata and controls

118 lines (94 loc) · 4.88 KB
Raw

Changes

3.5

  • Switching to a Snakemake profile method for setting resources.

3.4.3

  • Removed references to numpy primitives (np.int, np.float, etc.)
  • Updated submodules (pulled in fixes)
  • Index (#CHROM) type for svpoplib.ref.get_df_fai is str.

3.4.2

  • Documented flag rules
  • Automatically set missing IDs in input BED files for missing IDs
  • DipCall VCF parsing parameters
  • Split VCF parsing into batches for input that takes a long time to process in one job
  • Minor updates to support variant track output

3.4.1

  • Added dvpepper (PEPPER-Margin-DeepVariant) and SVIM input VCF parsers.
  • Added min_svlen and max_svlen options for VCF input parsers.

3.4.0

  • Merging and intersects bring SVs and indels together for the intersect and separate back out after.
    • Avoids mis-merging at SV/indel boundaries.
  • Added flag rules to run many related operations at once (for each sample in a sample list).
  • Subset chromosome option supports multiple chromosomes.

3.3.7

  • Added "pavbedhap" variant source for PAV input from each haplotype (not merged at the sample level).
  • Added filter pass option for parsing VCFs and controlling which variants are accepted based on the FILTER column.
  • Delly input VCF parser.
  • Fixed bug with upstream deletions (ALT=* was not ignored)

3.3.6

  • SNV format changed to "CHROM-POS-SVTYPE-REFALT" (eliminated dash between REF and ALT). All variant IDs can be split on "-" to obtain 4 fields (note that a ".n" may also be appended to distinguish multiple calls of the same type and at the same location).
  • Improved multi-allele VCF handling. Resulting tables now include "VCF_ALT_IDX" to match the alt genotypes each record was retrieved from. Each alternate allele will be separated into a separate record and assigned to samples where VCF_ALT_IDX is in GT.

3.3.5

  • Added threads to rules (supports --cores)
  • Fixed output format in rule hpref_merge_bed (was not gzipped)
  • Moved VCF parsing code to svpoplib (supports svpoplib use as an external library)
  • Input VCF parser assumes GT "1/." if the GT field is missing
  • Removed GT field for CuteSV (writes "./." causing the parser to drop variants)
  • Better support for writing VCF files with no variants (was causing crashes)
  • Can input dataframes to merging/intersect svpoplib routines (supports svpoplib use as an external library)

3.3.4

  • Fixed parsing Sniffles2 VCFs that are incorrectly formatted with "N" and SEQ in REF/ALT instead of symbolic ALTs

3.3.3

  • Bug in ID de-duplication code
  • Sorting RMSK annotations

3.3.2

  • Fixed bugs processing chromosome names containing a "." when de-duplicating variant IDs

3.3.1

  • Removed verbose output from merging (now optional)

3.3.0

  • Fixed pavlib functions PAV uses that cause numeric chromosome problems.

3.2.0

  • Set ply submodule version to 3.10
  • Alt-map retains CIGAR operations
  • Alt-DUP calls inner-variants (SV/indel INS/DEL & SNVs) from duplications using the alt-map CIGAR

3.1.0

  • CCRE annotations needed to be sorted.
  • VCF input support for DeepVariant and multiallelic sites.
  • Moved Sniffles and SVIM-asm input parsers to the VCF parser framework (no longer custom parsers, not needed).
  • Variant FASTA files get FAI.

3.0.0

  • Added altdup for remapping INS as DUPs (allows DUP version to be treated as a separate callset - i.e. merging, intersects, annotations)
  • Merging handles multi-allelic sites better.
  • Flexible merging parameter backend.
  • Revived VCF writer for standard BED files.
  • VCF: SVLEN header had the incorrect data type.

2.3.3

  • Dropped "expand" support for merging
  • Dropped "MERGE_AC" and "MERGE_AF" columns. These are not true AC and AF calculations without confident genotypes, which depends heavily on the input callset. A future version might consider GT if present.
  • Added "MERGE_N". Counts the number of samples supporting a variant.

2.3.2

  • Changed default merging parameters

2.3.1

  • Updated pipeline documentation

2.3.0

  • Added kanapy as a submodule

2.2.0

  • Added "match" option to nr strategy. Value is a comma-separated string where fields may be missing or empty to accept the default parameter (e.g. "match=0.75"). Fields:

    • SCORE_PROP [0.8]: Alignment score must be this proportion of the max where max is the score if all bases match (MATCH * max_length(A, B))
    • MATCH [2.0]: Alignment match score.
    • MISMATCH [-1.0]: Alignment mismatch score.
    • GAP-OPEN [-5]: Alignment gap-open score.
    • GAP-EXTEND [-0.5]: Alignment gap-extend score.
    • MAP-LIMIT [500000]: Use Jaccard distance if the largest of the two sequences is larger that this size. Similarity is still SCORE_PROP.
    • JACCARD-KMER [9]: Jaccard k-mer size.

  • Merge strategy "nrid" was removed. Replace with "nr:refalt" to do exact matches over position requiring REF and ALT to match.

  • Merge strategy "fam" was removed. This was an NR strategy that added columns for inheritance. This annotation should be done outside SV-Pop.