run_ephys_clustering.sh¶

This bash script runs several R and Python scripts in order to produce a set of stable clusters from sparse PCA electrophysiology values. The script takes a JSON file with the following parameters as an input:

Parameter	Description
`components_file`	Path to CSV file with sPCA components (input file)
`tau_file`	Path to file with cluster membership probabilities (output of run_ephys_clustering.r)
`labels_file`	Path to file with cluster labels (output of run_ephys_clustering.r)
`bic_file`	Path to file with Bayes information criterion values for different numbers of clusters (output of run_ephys_clustering.r)
`merge_info_file`	Path to JSON file with number of components after entropy-based merging (output of drcme.bin.run_post_r_merging)
`post_merge_tau_file`	Path to file with cluster membership probabilities after entropy-based merging (output of run_entropy_cumul_merges.r)
`post_merge_labels_file`	Path to file with cluster labels after entropy-based merging (output of run_entropy_cumul_merges.r)
`jaccard_file`	Path to file with Jaccard coefficient values after subset-based stability analysis (output of run_cluster_stability.r)
`etypes_file`	Path to file with stable e-type assignments (output of drcme.bin.run_merge_unstable_clusters)
`post_merge_proba_file`	Path to file with post-merging cluster membership probabilities (output of drcme.bin.run_merge_unstable_clusters)
`merge_unstable_info_file`	Path to file with merge sequence information (output of drcme.bin.run_merge_unstable_clusters)
`entropy_piecewise_components`	Number of components (2 or 3) for piecewise linear fit of entropy scores
`outliers`	List of outlier specimen IDs to exclude (use empty list if none)