Name | Modified | Size | Downloads / Week |
---|---|---|---|
Parent folder | |||
readme.txt | 2015-01-30 | 732 Bytes | |
10Fold_CrossValidation.r | 2015-01-30 | 9.5 kB | |
Figure_3.pdf | 2015-01-30 | 281.6 kB | |
Human_train.dat | 2015-01-30 | 954.0 kB | |
Totals: 4 Items | 1.2 MB | 3 |
Many users asked about how to generate the Two-graph ROC curve (i.e. panel D of Figure_3 in our CPAT paper: http://nar.oxfordjournals.org/content/41/6/e74.full) to determine the optimum cutoff. Here we provide the R code and the associated data table, so that uses could exactly reproduce Figure 3 in our paper. It is fairly easy to change the R code to accommodate their own data. Human_train.dat (20,000 rows includes 10,000 coding and 10,000 noncoding genes): Column-1: gene ID Column-2: mRNA size Column-3: ORF size Column-4: Fickett score Column-5: hexamer usage Column-6: Label (1: coding, 0: noncoding) Note: ROCR (http://cran.r-project.org/web/packages/ROCR/index.html) must be installed to run the above R code.