using TcGSA for nanostring data

[ShuyiGan]

Hi Boris! Hopefully I can still catch you here. I recently tried to download your TcGSA package to analyse a clinical trial dataset (nanostring of approximately 550 genes). I am just trying it out with 1 patient with 5 time points first but I keep getting an error

Error in TcGSA::TcGSA.LR(expr = Cohort3_RawCounts, gmt = Reactome.gmt, : 'expr' is not numeric. Don't know how to deal with nnon-numerical expressions.

I'm pretty new to R and wonder if this is a formatting problem but I can't fix it. All the data in the expr file is already numeric. Is there something that I'm missing?

thanks! Cohort3_RawCounts.txt DataMatrix.txt Reactome.gmt.txt

[borishejblum]

Hi Shuyi Gan,

First, thank you for your interest in our software.

Your error comes from the first column of Cohort3_RawCounts which is not a numeric value (genenames). This information should appear as row.names and not as a column in expr.

However, I do not think that TcGSA is suitable for your analysis because:

1. you have nanostring data, which are count data, while TcGSA was developped for continuous microarray data
2. You only observe 1 patient, while TcGSA models the heterogeneity between patients and therefore will not be able to properly estimate the associated random effects in your case.

For those reasons, I advise not to use TcGSA in your case.

Best, Boris