• All Files
Home/
Analysis Output/Output Files/Output Fusion Files/fusions.vcf.gz

Related articles

fusions.vcf.gz

This file contains information about detected gene fusion events.

The file is in VCF 4.1 format and compressed by gzip compression software. In addition to a format-compliant header, the file contains the following columns:

Column Heading

Description

CHROM

The chromosome on which the fusion occurs.

POS

The position of the gene fusion.

REF

The reference base preceding the fusion.

ALT

The alternative allele at the variant position.

ID

A unique identifier for the fusion event, formatted as: <Target ID>.<Target Number>.<A or B> for breakpoint A or B.

QUAL

Will always be "." (missing).

FILTER

Contains information on whether the variant passed quality filtering (PASS) or the name of the filter that caused it to fail quality filtering.

• Entries with scores ≥ FusionScoreCutoff (default value is 4.0) are marked as PASS and called Present.
• Entries with scores < FusionScoreCutoff and ≥ FusionEmitScoreCutoff are filtered as LowScore and called NoCall.
• Entries with scores < FusionEmitScoreCutoff (default value is 1.0) are not included in the report.
• The thresholds apply to both on- and off-targets.
• Targets are checked for OffTarget and LowScore. The rest will be marked as PASS.

INFO

Consists of the following sub-fields:

• SVTYPE—BND
• Two "SVTYPE=BND" records are created for each fusion call, one for each breakpoint)
• MATEID—<Target ID>:<B or A> for breakpoint A or B.
• READS—The target read count.
• SCORE—The score is calculated using the read count of the target and the total read count of the sample:

,

• Where totalreadcount is the number of reads mapped to all the targets in the manifest. The normalized read count is centered by 32.74276 and scaled by 254.8784 to produce a z-score-like score, where the two constants were learned from control samples without known fusions.
• The higher the score, the stronger the evidence of a fusion being present in the sample in question. Potential fusions with scores below 1 are not included in the fusions.csv file and potential fusions with scores lower than the 4 are filtered as LowScore.
• ANT—Annotation from the manifest for breakpoint A or B. Formatted as Gene|Transcript|Exon|Strand, where multiple exon numbers are joined by a comma (,). and strand is either + or -.

FORMAT

Consists of the following sub-fields:

• DP—The number of reads supporting this fusion.
• SC—The score of this fusion.