Skip to content

Compound Label #47

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Gautam8387 merged 3 commits into from
Jul 16, 2026
Merged

Compound Label #47

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
555e420
Add compound label decomposition and Hungarian cytescore_compound sco…
Gautam8387 Jul 14, 2026
4f1c224
wip
Gautam8387 Jul 14, 2026
a001446
wip
Gautam8387 Jul 14, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,8 +4,11 @@ tmp/
.env
.mypy_cache/
.pytest_cache/
.ruff_cache/
*.csv
*.owl
*.npz
*.npz.meta
*.json
cyteonto_results/
dist/
13 changes: 7 additions & 6 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,13 +2,14 @@

`cyteonto` compares two sets of cell type annotations against the [Cell Ontology (CL)](https://obofoundry.org/ontology/cl.html). Given label lists from a study author and one or more annotation algorithms, it:

1. Generates a structured description for every label with an LLM.
2. Embeds those descriptions with a configured embedding model.
3. Matches each embedding to the closest CL term.
4. Scores each author/algorithm pair using an ontology-aware similarity metric (default: a Gaussian kernel on the cosine similarity of the CL term embeddings).
5. Returns a tidy DataFrame with one row per `(algorithm, pair_index)`.
1. Decomposes mixture labels (doublets, mixed populations) into cell-type parts with an LLM when needed.
2. Generates a structured description for every label (or part) with an LLM.
3. Embeds those descriptions with a configured embedding model.
4. Matches each embedding to the closest CL term.
5. Scores each author/algorithm pair using an ontology-aware similarity metric (default: a Gaussian kernel on the cosine similarity of the CL term embeddings). Compound pairs use Hungarian bipartite matching with an optional coverage penalty when part counts differ.
6. Returns a tidy DataFrame with one row per `(algorithm, pair_index)`.

Updated ReadMe: [cyteonto/README.md](cyteonto/README.md). More documentation to follow!
Updated ReadMe: [cyteonto/README.md](cyteonto/README.md). Process flow and file layout: [docs/WORKFLOW.md](docs/WORKFLOW.md), [docs/FILE_MANAGEMENT.md](docs/FILE_MANAGEMENT.md).


## Modal Service
Expand Down
132 changes: 88 additions & 44 deletions cyteonto/README.md
View file
Open in desktop

Large diffs are not rendered by default.

2 changes: 2 additions & 0 deletions cyteonto/config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -78,8 +78,10 @@ class Config:
"author_label",
"algorithm_label",
"author_ontology_id",
"author_ontology_name",
"author_embedding_similarity",
"algorithm_ontology_id",
"algorithm_ontology_name",
"algorithm_embedding_similarity",
"cytescore_similarity",
"similarity_method",
Expand Down
Loading
Loading