Implement log1p transformation for cell count prior to PCA data.

Goal is to deal with extreme variance in cell count data via a log(x+1) transformation. Updated similarity metrics of all umap to "correlation" to better address data type found results. Also tested cosine similarity.
cns-iu · Aug 23, 2024 · 96f31ec · 96f31ec
1 parent 614e0b8
commit 96f31ec
Show file tree

Hide file tree

Showing 5 changed files with 74 additions and 60 deletions.
diff --git a/validations/output/umap/UMAP_Azimuth-PCA-Kmeans.png b/validations/output/umap/UMAP_Azimuth-PCA-Kmeans.png
diff --git a/validations/output/umap/UMAP_CellTypist-PCA-Kmeans.png b/validations/output/umap/UMAP_CellTypist-PCA-Kmeans.png
diff --git a/validations/output/umap/UMAP_PopV-PCA-Kmeans.png b/validations/output/umap/UMAP_PopV-PCA-Kmeans.png
diff --git a/validations/umap/.ipynb_checkpoints/HRA_UMAP_DatasetOrganCellType-checkpoint.ipynb b/validations/umap/.ipynb_checkpoints/HRA_UMAP_DatasetOrganCellType-checkpoint.ipynb
diff --git a/validations/umap/HRA_UMAP_DatasetOrganCellType.ipynb b/validations/umap/HRA_UMAP_DatasetOrganCellType.ipynb