API Reference¶
This is the API reference of DRCME. If you are interested in running standard analysis tasks, please look at the scripts reference page for more information.
drcme.ephys_morph_clustering
: Combined electrophysiology and morphology clustering¶
The drcme.ephys_morph_clustering
module contains functions for performing
joint clustering of electrophysiology and morphology data.
Functions¶
|
Calculate Jaccard coefficient |
|
Calculate co-clustering rates between clusters |
Determine consensus clusters from multiple variations |
|
Add Gaussian mixture model clustering results |
|
Add agglomerative clustering results with connectivity constraints |
|
Add agglomerative hierarchical clustering results |
|
Reassign samples to the best-matched clusters |
|
Add spectral clustering results |
|
Calculate Jaccard coefficients for subsampled clustering runs |
drcme.load_data
: Data handling of IPFX outputs¶
The drcme.load_data
module contains functions for loading
electrophysiology feature vectors processed by the IPFX package, as well as sPCA parameter files.
In particular, it loads in HDF5-format files containing feature vectors
processed by the run_feature_vector_extraction script
.
Functions¶
|
Load an sPCA parameters file |
|
Load dictionary for sPCA processing from HDF5 file |
drcme.post_gmm_merging
: Merging Gaussian mixture model components¶
The drcme.post_gmm_merging
module contains functions for merging
Gaussian mixture model components together based on an entropy criterion
as in Baudry et al. (2010).
Functions¶
|
Merge clusters by entropy criterion and piecewise fit |
Merge set of specified clusters by entropy criterion |
|
|
Fit entropy vs cumulative merge number with linear piecewise function |
Reorder cluster labels by similarity of centroids |
drcme.prediction
: Type label prediction¶
The drcme.prediction
module contains wrapper functions for random
forest classification.
Functions¶
|
Predict labels for test_df by random forest classification |
drcme.spca
: Sparse principal component analysis¶
The drcme.spca
module contains functions for performing
sparse principal component analysis.
Functions¶
|
Combine and z-score individual data set sPCs into single matrix |
Recover mean and standard deviation of z-scored sPCs |
|
|
Select data sets and indices defined by spca_params |
|
Compute sPCA for multiple data sets with specified parameters |
|
Transform and z-score new data with existing loadings |
|
Perform sparse principal component analysis |
drcme.tsne
: t-SNE helper functions¶
The drcme.tsne
module contains wrapper functions for
applying t-SNE to the electrophysiology and morphology data.
Functions¶
|
Perform t-SNE on two data sets |
|
Perform t-SNE on electrophysiology and morphology data sets |