New: Check our last work on genetic mutation predictions in PDA with pathomic and transcriptomic data! -> https://github.com/FraBer95/MultimodalMutationsPDAC
This is the repository of the "A time-dependent explainable radiomic analysis from the multi-omic cohort of CPTAC-Pancreatic Ductal Adenocarcinoma" paper.
The code is written in R and Python and it is composed by the following parts:
- data folder, that should contains all data in a csv format and the pyradiomics configuration .yaml file in
- python_scripts folder, containing all pre-processing operations for merging different data, extraction and validation of the radiomic signature
- R_scripts folder, containing the feature selection in UV, feature dicotomizing and survival models training and test
Preprocessing part: Data Reading and formatting, dropping of non relevant features, categorization, statistical analysis and feature selection. Analysis part: Survival Curves, Models Training and Validation, XAI(t). The python libraries requested are:
-pandas, torch, pycox, reticulate, numpy, seaborn, scipy, statsmodels, scikit-learn
The R libraries requested are:
-survival, surviminer, survex, randomForestSRC, gbm, survivalsvm, ggsurvfit, ggplot2, pec, caret, SurvMetrics, mlr3proba, mlr3extralearners, mlr3pipelines, survivalmodels, reticulate