How guangyuwanglab2021/loki works
Loki is a Python toolkit for spatial-transcriptomics and H&E analysis built around five documented capabilities: tissue alignment, cell-type decomposition, tissue annotation, ST gene-expression prediction, and image–transcriptomics retrieval. A typical user prepares data and embeddings, invokes the relevant analysis module, then inspects results with plotting utilities. The README documents these capabilities and module imports; sampled source confirms the modules’ individual computations. It does not establish cross-module orchestration, so no inter-module workflow edges are inferred.
Main entry points: preprocess.py Data preprocessing · utils.py Model and embeddings · align.py Tissue alignment · annotate.py Bulk annotation
Components 10
Data preparation
- Data preprocessing
src/loki/preprocess.py
Feature encoding
- Model and embeddings
src/loki/utils.py
Analysis capabilities
- Tissue alignment
src/loki/align.py - Bulk annotation
src/loki/annotate.py - Marker annotation
src/loki/annotate.py - Cell decomposition
src/loki/decompose.py - Image ST retrieval
src/loki/retrieve.py - Gene expression prediction
src/loki/predex.py
Results and visualization
- Result plotting
src/loki/plot.py
Other components
- Loki user
Connections 9
- Loki user → Data preprocessing: prepares data
README.md - Loki user → Model and embeddings: encodes features
README.md - Loki user → Tissue alignment: aligns tissue
README.md - Loki user → Bulk annotation: annotates tissue
README.md - Loki user → Marker annotation: annotates tissue
README.md - Loki user → Cell decomposition: decomposes cell types
README.md - Loki user → Image ST retrieval: retrieves matches
README.md - Loki user → Gene expression prediction: predicts expression
README.md - Loki user → Result plotting: visualizes results
README.md