|Meta prediction of protein crystallization propensity|
Three dataset are used:
- TEST144 was developed by Overton and colleagues (2008) (originally this set was called TEST) and can be downloaded from http://compbio.dundee.ac.uk/xtal/
- The TRAINING and TEST500 dataset can be downloaded from TRAINING set link and TEST500 set link
The datasets are in a tab-separated text file format with three columns that correspond to sequence id (from either TargetDBTargetDB or PepcDB), annotation of the class label (crystallizable or noncrystallizable), and sequence.