Track Interpretation
4 properties
What the data in a genomic annotation file represents. Interpretation has two distinct halves — an algorithmic concept (the process/output, e.g. "peaks") and a biological concept (the entity, e.g. "TF binding site"). These draw on the DataType and FeatureType vocabularies respectively. Corresponds to FGA-WG issue #2 (data interpretation). This schema records what the data is; for what was done to it (read selection, QC filtering), see Track Provenance.
output_type
The algorithmic interpretation: the computational process or output the file represents (e.g. peaks, signal, alignment, count matrix). Draws on the DataType vocabulary.
feature_type
The biological interpretation: the biological entity or concept the data represents (e.g. TF binding site, enhancer, CpG island). Draws on the FeatureType vocabulary.
strand
The strand orientation the file represents. Factored out of compound DataType terms (e.g. "minus strand signal"). Absent means unstranded/unspecified.
haplotype_resolution
The haplotype resolution the file represents: whether content is resolved to individual alleles (allele_specific), partitioned per parental haplotype (haplotype_specific), or phase-resolved without per-haplotype splitting (phased). A content facet factored out of compound DataType/FeatureType terms (e.g. "haplotype-specific contact matrix", "phased variant calls"). Absent means haplotype-collapsed / not phase-resolved.
The strand and
haplotype_resolution slots are
content facets, factored out of compound DataType / FeatureType terms
(e.g. "minus strand signal" → output_type: signal, strand: minus;
"haplotype-specific contact matrix" → output_type: contact matrix,
haplotype_resolution: haplotype_specific); the
ENCODE round-trip is preserved in mappings/facet_decomposition.tsv.
Both are intrinsic content, so they stay here. The processing facets
read_multiplicity and
filter_status describe what
was done to the data and now live on
Track Provenance.