Research article
Assessing the druggability of protein-protein interactions by a supervised machine-learning method
-
* Corresponding author: Nobuyoshi Sugaya sugaya@pharmadesign.co.jp
Drug Discovery Department, Research & Development Division, PharmaDesign, Inc, Hatchobori 2-19-8, Chuo-ku, Tokyo 104-0032, Japan
BMC Bioinformatics 2009, 10:263 doi:10.1186/1471-2105-10-263
Published: 25 August 2009Additional files
Additional file 1:
Supplementary tables. Table S1 lists the positive set PPIs in more details than Table 1 in the text. Table S2 lists the top 10 attributes by F-score. Table S3 lists the potentially-druggable PPIs predicted by the SVM-based method in more details than Table 4 in the text. The 69 instances (42 PPIs) listed have the druggability scores of >9,000 by the SVM models using all attributes and >6,500 by the models using the top 10 attributes. Table S4 lists 9 small chemicals showing similarities to the hot spots of the SMAD4/SKI complex. File format, PDF.
Format: PDF Size: 214KB Download file
This file can be viewed with: Adobe Acrobat Reader
Additional file 2:
Positive instances. The 98 instances are listed with their attributes. Attribute numbers correspond to those in Table 2.
Format: XLS Size: 111KB Download file
This file can be viewed with: Microsoft Excel Viewer
Additional file 3:
Test instances. The 10,915 instances are listed with their attributes. Attribute numbers correspond to those in Table 2.
Format: XLS Size: 8.9MB Download file
This file can be viewed with: Microsoft Excel Viewer
Additional file 4:
Supplementary methods. Definition and calculation methods of the PPI attributes.
Format: PDF Size: 727KB Download file
This file can be viewed with: Adobe Acrobat Reader
Additional file 5:
Assessment of the druggability of the PPIs. For each positive or test instance, the druggability is assessed by the druggability scores (the numbers of times an instance was judged to be positive more than 9,000 times with the SVM models using all attributes and more than 6,500 times with the SVM models using the top 10 attributes).
Format: XLS Size: 1.5MB Download file
This file can be viewed with: Microsoft Excel Viewer
Additional file 6:
Figure S3. Discovery of small ligands showing similarities to the hot spots of the SMAD4/SKI complex.
Format: PDF Size: 2.1MB Download file
This file can be viewed with: Adobe Acrobat Reader
Additional file 7:
Ligand-binding pockets of the positive PPIs. Coordinate data of the ligand-binding pockets of the positive PPIs.
Format: TXT Size: 149KB Download file