Anika preprocess analysis

[anikaliu]

DoRothEA and PROGENy preprocessing.

Added/Updated the necessary files: -DoRothEA regulon -PROGENy matrix -Pathway-representative nodes

Added vignette+ functions to -run DoRothEA/PROGENy from t-values -generate corresponing CARNIVAL input files

[ptrairatphisan]

Please check the following: 1) Comment the hard coded "setwd("~/Documents/GitHub/CARNIVAL/")" on the DEG_preprocessing script as the path to the CARNIVAL folder might be differed in each machine 2) Dependency of addition packages upon running GEX processing script (dplyr is needed for the pipe %>% operation and viper is needed for calculate TF scores) -> load these libraries before calling the function 3) [major] The nested function "map_df" didn't work on the machine I test (error code: Error in map_df(., .f = function(i) { : could not find function "map_df"). Please check if additional package is required to load such nested function.

These issues should be resolved before I could further proceed with code evaluation. In addition, it should be noted that the list of DoRothEA regulons provided here came from a sub-selected set of "Best" regulons and not the whole set of regulons. This information needs to also be documented in the manuscript.

[anikaliu]

Added dependencies. Will add as example in the next few days.

[ptrairatphisan]

It appears that the nested function requires the package 'tidyverse' to be loaded. Now the function run_dorothea.R runs fine.

Nevertheless, the function genesymbol2uniprot.R which is called in the next step return the following error: "Error in names(x) <- value : 'names' attribute [296] must be the same length as the vector [3]". This issue came from the following line of code: "colnames(df)<-df$uniprotids_for_mapping" where df only has 3 columns but the name-replacing vector is of length 296. Did you mean to map to "rownames" instead of "colnames" here?

In addition, the variable "TF_activities" is missing as an input in the function generate_measfile.R in the next step. Please check the consistency of variable names and supply the required input accordingly.

[anikaliu]

This is not the up to date version you are looking at. This was already corrected, and also the variable manes are different in the current version.

[ptrairatphisan]

Yes, now the issues mentioned above were resolved in the updated version. The script returns no error anymore.

Still one last minor thing before the final acceptance: PROGENy scores in the score file should be separated by tab (and not by space per default setting) to be compatible with the input format in CARNIVAL. Please update the DEG_preprocessing.R script/vignette accordingly.

[anikaliu]

Still one last minor thing before the final acceptance: PROGENy scores in the score file should be separated by tab (and not by space per default setting) to be compatible with the input format in CARNIVAL. Please update the DEG_preprocessing.R script/vignette accordingly.

Done!

[ptrairatphisan]

Final test passed. CARNIVAL measurement and score input files were correctly generated from the DEG processing pipeline and are compatible with the CARNIVAL pipeline. CARNIVAL results/networks were also correctly generated from the driver script for example 2 (SBVimprover - EGF).