GeorgescuC
commented
2 years ago Hi @JesseRop ,
There is in theory no reason why it should not work if a few conditions are met. You will want to check that:
- the CNVs you are interested in are large enough, as in, they cover multiple expressed genes. Single cell data being very noisy, CNVs of a single gene are not detectable by this method.
- the proximity of genes part of a CNV is conserved, meaning a large enough piece of DNA is duplicated/deleted as a whole and not dispersed throughout the rest of the genome.
These limitations are due to one of the core aspects of infercnv that uses the proximity of genes on the genome to identify the larger scale CNVs by smoothing the expression of genes over a sliding window.
As for the run settings, since you are looking at a haploid organism, if you want to run the HMM, you should run it in i3 mode rather than the default i6 as the i6 model assumes a base diploidy.
Regards, Christophe.
Dear Developers,
Thank you very much for this great tool. I work on plasmodium parasites and have scRNAseq data for different strains of these which are known to have CNVs. Do you know if I can use your tool for this and how i'd go about it.