Skip to content

kristiyanto/DREAM9.5

Repository files navigation

                	DREAM CHALLENGES SUBMISSION (CHALLENGE 2.B)

                  		UNIVERSITY OF WASHINGTON TACOMA
                  		
                                  SUMMER - 2015

Predicting Discontinuation of Docetaxel Treatment for Metastatic Castration-Resistant Prostate Cancer (mCRPC).

  1. INTRODUCTION This is a continuation of Team Yoda submission. The contributor of this reboot is Daniel Kristiyanto, Kevin Andersonn, and Ka Yee Yeung served as the PI.

  2. FILE LIST MAIN.R : Contains the main R script to analize and generate prediction. inc_functions.R : Contains functions such as data split, etc, required by Main Script. inc_datacleanup.R : Contains function to perform data cleansing and pre-processes score.R : Contains the scoring script to compute AUC as provided by Synapse CoreTable_training.csv : Contains the training data as provided by Synapse CoreTable_validation.csv : Contains the Validation data as provided by Synapse. CoreTable_leaderboard.csv : Contains the Validation data as provided by Synapse. This will be concatenated with CoreTable_validation.csv

    README.txt : this file.

  3. REQUIRED R PACKAGES (FSelector) (e1071) (timeROC) (caret) (unbalanced) (impute) By Bioconductor

  4. VARIABLES CV : Numeric vector to perform thec cross-fold validation. Each folds takes about 25 minutes to run. Sample: CV = c(1:10) STUDY : Numeric vector ranged of 1 to 3; Script itterates through every study with following target (and use the other two as training data): 1. ACCENT2, 2. EFC6546, 3. CELGENE 4. ALL COMBINED. Sample: STUDY = c(1,2,3)

  5. DATAFRAMES Once run completed, these variables should be available: core_table_training : Dataframe, Original core training data (Raw) core_table_validation : Dataframe, Original core testing data (Raw) table.for.model : Dataframe, Cleaned and augmented core training data table.for.validation : Dataframe, Cleaned and augmented core testing data curr.training.data : Dataframe, Subset of current fold training data curr.testing.data : Dataframe, Subset of current fold testing data SCORING.TABLE : Dataframe, Scoring results OUTPUT.TABLE : Dataframe, Subset of current testing data with predicted probability FINAL.TABLE : Dataframe, Subset of core testing data with predicted probability

  6. OUTPUT Within each itteration, following files are generated: SCORE-[STUDYNAME]-[AUC].csv : Contains the scoring information. SUBSETTEST-[STUDYNAME]-[AUC].csv : Contains the Subset of the testing data used for the final prediction. VALIDATION-[AUC].csv : Contains the prediction of the validation data as required by Synapse.

About

Predicting Discontinuation of Docetaxel Treatment for Metastatic Castration-Resistant Prostate Cancer (mCRPC)

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages