Entering edit mode
6 days ago
adR
▴
120
Hi everybody,
I tried to use Kassandra, a deconvolution algorithm that estimates the proportion of cells in the Tumor microenvironment (TME). My RNA seq data is in a matrix format, where genes are in the rows, and samples are in the column. Values are TPM, also the FPKM on the other dataset. I trained the model as they said in the Tutorial LINK
I am not 100% sure if I understand their requirement and I’m wondering if Kassandra would work with datasets where the alignment was performed using tools such as STAR, Bowtie2, etc, other than Kallisto. If so, could you explain why that is the case?
kallisto is one of many valid ways to get a count matrix, so I doubt that is the only valid way. They just like that tool, so they seem to use it. bowtie2 is not splice-aware so it is not valid here.