This PR adds a kallisto workflow. It runs kallisto quant on all samples and then runs sleuth for differential expression.
sleuth is an R package, and the sleuth rule is an example of running a customized R script. So there is a new helpers.rscript function that writes and runs the filled-in-template. This is because I like having the script available for later debugging/tweaking. I'm still trying to figure out it it's worth it to put sleuth into a wrapper or not.
Also note that the model for sleuth is specified in the config file.
kallisto is kind of annoying because if you want to use the downstream sleuth package, you need to have particularly-named files. See the kallisto quant readme for details. This kind of breaks the existing SampleHandler class system, hence the double-brackets in the patterns.
Not sure I'm happy with the output organization or with the R script handling, but it works. Please take a look and see what you think. Next step will be reporting.
This PR adds a kallisto workflow. It runs
kallisto quanton all samples and then runssleuthfor differential expression.sleuthis an R package, and the sleuth rule is an example of running a customized R script. So there is a newhelpers.rscriptfunction that writes and runs the filled-in-template. This is because I like having the script available for later debugging/tweaking. I'm still trying to figure out it it's worth it to put sleuth into a wrapper or not.Also note that the model for sleuth is specified in the config file.
kallisto is kind of annoying because if you want to use the downstream sleuth package, you need to have particularly-named files. See the kallisto quant readme for details. This kind of breaks the existing SampleHandler class system, hence the double-brackets in the patterns.
Not sure I'm happy with the output organization or with the R script handling, but it works. Please take a look and see what you think. Next step will be reporting.
Test it with: