I am working on a set of single-nuclei sequencing samples. Does someone know why I might be getting a bimodal distribution in one of my samples (both in features/cell and reads/cell)? The modes are very different (1000 genes vs 150 genes). The other samples almost exclusively contain the "small mode" (i.e. ~150 gene mode). I have tried doublet removal which preferentially removes the big mode, but not enough to help. I also tried finding dead/dying cells based on mitochondrial fraction, but (a) it didn't help much and (b) I'm not even sure if that is appropriate for single nuclear samples rather than single cell samples.
Does anyone know what this may be? Is there any way for me to confidently ascertain if this represents real biology?
also, how can I flag disintegrating nuclei in single nuclei sequencing data?