ORFcompare

ORFcompare is a companion tool to ORFanage that compares CDS annotations between query and template transcripts. While ORFanage is designed to annotate new ORFs, ORFcompare evaluates and quantifies the differences between existing CDS annotations in two GTF/GFF files.

This tool is particularly useful for:

  1. Validating ORFanage results against known annotations

  2. Comparing CDS annotations from different sources

  3. Quantifying frame preservation between transcript isoforms

  4. Identifying transcripts with matching or divergent coding regions

Basic Usage

At minimum, ORFcompare requires a query file, a template file, and an output file:

$ orfcompare --query query.gtf --template template.gtf --output comparison.tsv

To include start and stop codon information, provide a reference genome:

$ orfcompare --reference genome.fa --query query.gtf --template template.gtf --output comparison.tsv

Output Format

ORFcompare outputs a tab-separated file containing metrics for each query/template pair comparison. For each query transcript that overlaps with a template transcript, the following metrics are computed:

  • Length metrics: CDS lengths for both query and template

  • Overlap metrics: Matching, in-frame, out-of-frame, extra, and missing bases

  • Percent identity metrics: LPI, MLPI, and ILPI scores

  • Codon information: Start and stop codons for both transcripts (requires --reference)

For detailed column descriptions, please refer to the ORFcompare Stats Output section in File Formats.

Description of Options

--query

Path to the GTF/GFF file containing query transcripts with CDS annotations to be compared.

--template

Path to the GTF/GFF file containing template (reference) transcripts with CDS annotations.

--output

Path to the output TSV file where comparison results will be written.

--reference

Path to the reference genome in FASTA format. When provided, enables extraction of actual start and stop codon amino acids for both query and template transcripts. This allows verification of proper translation initiation (M for methionine) and termination (. for stop codon).

--threads

Number of threads to use for parallel processing. Similar to ORFanage, transcripts are grouped by coordinate overlap or gene ID and processed independently.

--use_id

When enabled, transcripts are grouped by gene ID rather than coordinate overlap. This is useful when gene IDs are consistent between query and template files.

Interpreting Results

The key metrics to evaluate CDS similarity are:

ILPI (In-frame Length Percent Identity)

The most important metric for coding sequence comparison. High ILPI (>90%) indicates the query and template share most of their coding sequence in the same reading frame.

Start/Stop Codon Match

  • M for start codon indicates canonical translation initiation

  • . for stop codon indicates proper termination

  • Other amino acid letters indicate incomplete or alternative codons

len_extra / len_missing

These values quantify how the query CDS differs from the template:

  • High len_extra: Query has additional coding sequence not in template

  • High len_missing: Query is missing coding sequence present in template

Example Analysis

Compare ORFanage output against the original reference annotation:

$ orfcompare --reference genome.fa \
             --query orfanage_output.gtf \
             --template original_annotation.gtf \
             --output validation.tsv

Then filter for high-confidence matches:

# Find transcripts with >95% in-frame identity
awk -F'\t' '$12 > 95' validation.tsv > high_confidence.tsv

# Find transcripts with matching start codons
awk -F'\t' '$13 == "M" && $14 == "M"' validation.tsv > matching_starts.tsv