sgkit
sgkit copied to clipboard
Document best practice of windowing after filtering varaints
sgkit stores genomic windows as the indices of their first and last variants within each window. This means that filtering variants (i.e. dropping some variants from the dataset) will invalidate the indices in window_start and window_stop. This should be documented and users encouraged to generate genomic windows after filtering any variants.
Related discussion: https://github.com/pystatgen/sgkit/discussions/786#discussioncomment-1936669