Documentation updated: 04/23/2020

See also https://ftp.ncbi.nlm.nih.gov/pub/dbVar/sandbox/sv_datasets/nonredundant/release_notes/NR_stats.latest.txt

FTP Directory: nonredundant
Last modified: Oct 30, 2023
File types: bed, bedpe, tsv

All files are available in bed, bedpe, and tsv formats:

https://ftp.ncbi.nlm.nih.gov/pub/dbVar/sandbox/sv_datasets/nonredundant/deletions/ https://ftp.ncbi.nlm.nih.gov/pub/dbVar/sandbox/sv_datasets/nonredundant/insertions/ https://ftp.ncbi.nlm.nih.gov/pub/dbVar/sandbox/sv_datasets/nonredundant/duplications/

• Sets of "non-redundant structural variations" (NR SVs) derived from dbVar are available via FTP as tab delimited files by assembly, GRCh37 & GRCh38, and type of variant.

• Non-redundant refers to variant coordinates, i.e. chr, outermost start, and outermost stop. Please note: the non-redundant coordinates are based strictly on exact overlap of coordinates, not on partial overlaps.

• Other features of NR SV files:

  • variant calls are from germline samples only (no somatic)
  • placements are "BestAvailable" on the assembly (guarantees no duplicate placements for a variant)
  • placements are on finished chromosomes only (not on NT_ or NW_ contigs)
  • placements are 1-based in the .tsv files
  • placements are zero-based start and 1-based stop in .bed and .bedpe files
  • insertion_length is set to sequence length if the sequence was submitted to dbVar without a specific insertion_length
  • insertions submitted to dbVar without insertion_length or submitted sequence are not included in the NR files

• Other files based on NR SV files:

  • NR SV files annotated with overlapping ACMG genes
  • NR SV files in .bed format
  • NR SV files in .bedpe format

Please note:

• * = NR_SV TSV fields 14 and 15 are in nr_insertion.tsv files only
• NR_SV_id = chr_outermost_start_outermost_stop_type where type is del, dup, or ins
• bin_size = small (length < 50 bp), medium (< 1000000), large (>= 1000000). Length = outermost_stop - outermost_start + 1.
• In all cases, bedpe columns 4 through 6, and 8 through 10, are populated with default values per the bedpe specification
• The bed and bedpe specifications are found here: https://bedtools.readthedocs.io/en/latest/content/general-usage.html

• The fields type, method, analysis, platform, variant, study, clinical_significance, clinvar_accession, and gene may contain multiple values.
• Each of the values is associated with one or more calls found in the variant field.
• The values in the variant field are "dbVar call accessions".

• only insertion SVs have minimum_insertion_length and maximum_insertion_length fields

• Placements in bed files are zero-based start and one-based stop
• name is comprised of chromosome, outermost_start, outermost_stop, and type (del, dup, or ins)
• NR SV .bed files may be used with a variety of tools as shown in the tutorial:

https://github.com/ncbi/dbvar/blob/master/Structural_Variant_Sets/Nonredundant_Structural_Variants/ToolGuide.md

• Placements in bedpe files are zero-based start and one-based stop
• bedpe files are normally used for disjointed genomic sequences
• the NR bedpe files do not contain disjointed genomic sequences, and instead use default values for chrom2, start2 and end2
• the NR bedpe files use the optional fields starting at field 11 to hold: chr, outermost_start, outermost_stop, name, and all the additional fields which are in the NR SV files

• ACMG files are provided as a "proof of principle" for a "use case" for the NR SV files.
• ACMG files are based on region/gene overlaps, and are missing a few call/gene overlaps in the cases where the parent variant region does not include all of the variant call placement and the region does not overlap the gene or is upstream or downstream of the gene.
• Placements in the ACMG files do not account for confidence intervals, even though the overlaps reported in the file were determined using confidence intervals. This results in a few of the overlaps that are not supported by the placements as reported in the file.

For information on ACMG genes please see: https://www.ncbi.nlm.nih.gov/clinvar/docs/acmg/

Please see README files for deletions, insertions, and duplications for example records and additional details.

• Deletions: https://github.com/ncbi/dbvar/blob/master/Structural_Variant_Sets/Nonredundant_Structural_Variants/Deletions/README.md
• Insertions: https://github.com/ncbi/dbvar/blob/master/Structural_Variant_Sets/Nonredundant_Structural_Variants/Insertions/README.md
• Duplications: https://github.com/ncbi/dbvar/blob/master/Structural_Variant_Sets/Nonredundant_Structural_Variants/Duplications/README.md

The algorithm makes use of previously existing scripts.

Input files are generated from the dbVar database with tab separated values and contain SVs by assembly, type, and other relevant fields.

Selected type files are grouped into "aggregated type files" as specified above, by chr.

The "aggregated type files" are converted into XML records containing all the neccessary fields required by the nr process.

The XML is then parsed to generate SV records with coordinates, type, method, analysis, platform, insertion_length, SV accession and study.

The SV records are then processed to generate NR SV tab-separated value (tsv) files by assembly and type, as described above, e.g.

• GRCh38.nr_deletions.tsv.gz

Examples of using the NR files with various tools and browsers can be found in: https://github.com/ncbi/dbvar/blob/master/Structural_Variant_Sets/Nonredundant_Structural_Variants/ToolGuide.md

• Please email dbvar@ncbi.nlm.nih.gov or create an issue on this GitHub page.

Thanks for your interest in the dbVar human "non-redundant structural variations" (NR SVs) data files from NCBI.

Please check back soon for further updates.