Skip to content

extraction of data embeddings from deep learning model layers; computation of embedding distance and visualization with umap/tsne

Notifications You must be signed in to change notification settings

ENCODE-AWG/locusselect

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

83 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

locusselect

  • extraction of data embeddings from deep learning model layers;

  • computation of embedding distance for inputs;

  • clustering and visualization of embeddings with umap/tsne

See examples/example.sh for usage examples.

 compute_embeddings \
                    --input_bed_file optimal_peak.narrowPeak.gz \
                    --weights 13kb_context_3b_prediction_dnase.hdf5 \
                    --json k562_dnase_profile_arch.json \
                    --ref_fasta /mnt/data/annotations/by_release/hg38/GRCh38_no_alt_analysis_set_GCA_000001405.15.fasta \
                    --center_on_summit \
                    --flank 6500 \
                    --output_npz_file k562_dnase_profile_embeddings_layer_-2.npz \
                    --embedding_layer -2 \
                    --threads 40

About

extraction of data embeddings from deep learning model layers; computation of embedding distance and visualization with umap/tsne

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 99.9%
  • Other 0.1%