This github repository provides codes for generating gene quantifications for pseudobulks.
Note 1: Many "count" values are not integers, as all of the ENCODE sc/snRNA-seq parse data was processed with this command: https://github.com/detrout/woldrnaseq-wrappers/blob/main/star_solo_splitseq/wrapper.py#L70
The gene_model is defined by the above function. Return "GeneFull_Ex50pAS" if config['include_intron'] else "Gene"def get_gene_model(config): where include_intron was true for the single nucleus data.
When running STAR, the filename is the gene mode. So if it was run as single nucleus the directory name will be GeneFull_Ex50pAS and should have floating point values.
All the parse data was run with GeneFull_Ex50pAS. Definition for the GeneFull_Ex50pAS can be referred to https://github.com/alexdobin/STAR/blob/b1edc1208d91a53bf40ebae8669f71d50b994851/source/parametersDefault#L863