[alpae]

Fixes #40 (part of #43).

1. Extant genes are now ordered by genomic coordinates.
The extant synteny graphs linked consecutive genes in GFF file order, but GFF files are not necessarily sorted. Genes are now sorted by contig (contigs keep the order of their first appearance in the GFF), start and end before they are linked.

2. The orientation of extant edges uses the correct strand.
It used the strand of the previous gene in the GFF (gene_keys[i-1]), which is a skipped gene if a gene missing from the orthoXML lies in between. It now uses the strand of the previous linked gene.

Effect on results

This changes the inferred ancestral graphs whenever a GFF is not sorted, so the version is bumped to 0.1.11.

The GFFs in test_data are not sorted. In CALJA.gff, for example, all 1,110 genes are on chromosome 12, but stretches near 3, 28, 43, 63, 71 and 106 Mb are interleaved: 975 of the 1,109 steps between consecutive gene lines jump backwards or more than 1 Mb forwards (9 gaps > 1 Mb in sorted order). The first gene line, ENSCJAG00000042835.2 at 12:28,070,673, was followed by a gene at 2,976,794; its real neighbors are 3, 5, 8, 11, … lines further down, and its upstream neighbors are at the end of the file.

Ancestral edges summed over all levels, test_data run with --orient_edges:

	main	this PR	shared
bottom-up	10,580	7,313	5,291
top-down	6,059	6,189	4,451
linearized	6,054	6,081	4,419

About 27% of the linearized ancestral edges change. The top-down phase now keeps 85% of the bottom-up edges, compared to 57% before, as far fewer spurious extant adjacencies are propagated. The ancestral edge checked by test_known_ancestral_adjacency is still inferred.

Not changed

The parser still assigns CDS lines to the gene line directly above them in the file (the Parent attribute is not used), so GFFs that do not list each gene's CDSs right after it are still misread. For nested or overlapping genes, sorting by start places a nested gene right after the gene containing it.

Tests: removed the xfail markers of test_unsorted_gff, test_interleaved_contigs and test_orientation_ignores_skipped_gene.

[Copilot]

🟢 Approval recommended

The focused changes address the reported defects with regression coverage and no identified blocking issues.

No open issues 🎉

What changed in this PR

Fixes extant synteny graph construction by using genomic order and the correct linked gene’s strand.

Changes:

• Sort genes by contig and coordinates before linking them.
• Use the previous linked gene’s strand for edge orientation.
• Enable three regression tests and bump the version to 0.1.11.

File	Description
uv.lock	Updates the locked project version.
tests/​test_extant_graphs.py	Removes expected-failure markers from three regression tests.
pyproject.toml	Bumps the project version to 0.1.11.
edgehog/​init_extant_synteny_graphs.py	Corrects gene ordering and edge orientation.

_{Review effort: Balanced}