`all_cba.KLEAT.on-target-cleaned.filtered.clustered.csv.gz` is the clean version of all cleavages used in the analysis after filtering and clustering. This is likely the version you would like to use. `all_cba.KLEAT.csv.gz` is raw cleavage site predictions concatenated from KLEAT outputs of all samples from the pipeline https://github.com/bcgsc/tasrkleat. If you would like to do the cleaning and filtering in a different way, you could use this version.