GitHub - HuangLabUMN/oncoPredict: An R package for drug response prediction and drug-gene association prediction. (original) (raw)

oncoPredict

Note, for questions on oncoPredict, pleast contact us at rshuang at umn.edu . The email associated with this account is not regularly checked.


Please Note the Following Bugs We are Aware of:

Selecting the batch correction option "qn" (microarray vs RNA-seq) will give an error. (caused by line 167 in CALCPHENOTYPE.R seems to lose row names)
Would recommend using normalizeQuantiles from limma package before running the models and then choosing 'none'.

Also, there is another option for microarray to RNA-Seq for batch correction, the 'standardize' option. This is currently not documented, but it works to calculate z-scores on the microarray and RNA-Seq data for integration.

Finally, oncoPredict currently expects a matrix and so it breaks if you try to build a model with just one variable. (E.g. if you were only interested in the lapatinib response, you'd have to run lapatinib and some other drug still. Not the biggest deal, but something to be aware of.


(Predict Response from Expression Data and Identify Cell line/Clinical Targets and Trends)

Additional details about this package can be found in our publication oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data

An R package for drug response prediction and drug-gene association prediction. The prepared GDSC and CTRP matrices for the calcPhenotype() are located in the oncoPredict OSF.

R

vignettes

man

NAMESPACE

DESCRIPTION

Figure 1.

Flowchart displaying the 3 primary functionalities available through oncoPredict (calcPhenotype, GLDS, IDWAS) as well as the files generated from each function and parameters. Functions and files generated are bold.

Figure 1_ Overview of pRRophetic_plus (a flow chart or similar diagram to highlight the package’s abilities)   (2)