Update run_filter.py

[deevdevil88]

There was an error in the for loop, the same variable wasnt being used across, leading to a mistake in the ncells counts per modality. as a stored mod variable from the previous for loop used durng filtering was being used instead. This leads to differences in cell numbers per sample_id per modality in the preprocess workflow output files output_prefix_cell_counts.csv and output_prefix_filtered_data.tsv. i wanted to thank @JHYSiu for finding this bug and flagging it. i have tested the change interactively in jupyter lab, and the ncell numbers per modality now match up with the preprocess object cel numbers.

Best,
Devika