Applicability to Cancer Samples

[DarioS]

Have you considered the applicability of Milo to cancer cell data? I notice in the journal article a variety of scenarios (e.g. cirrhotic liver, thymus), but not cancer. I suppose that it is because one person's cancer clone is quite different to another's. For example,

Epithelial cells, fibroblasts, endothelial cells, lymphocytes, and myeloid cell states contained a finite number of minor cell states, which were also reproducibly found in different patients. This was in stark contrast to the tumor cell states from the same group of patients, which were largely patient-specific.

Source: CXCL9:SPP1 Macrophage Polarity Identifies a Network of Cellular Programs that Control Human Cancers, Science.

I have seen other researchers using it on cancer cell data, but I think it is a misuse of MiloR because it often ends up as one neighbourhood group being comprised of one patient's cells. Is it perhaps worth an explanatory paragraph in vignette?

[MikeDMorgan]

Hi @DarioS Milo is a general tool for differential abundance analysis. There could be many reasons for many tumour cell states being patient specific, and these could be biological or technical reasons. There is no reason why Milo can't be used for analysing cancer data as long as the assumptions are reasonably met. This is not specific to cancer data, but any data set that an analyst wishes to apply Milo to.