This page contains the data for the real-valued drug-target binding affinity prediction experiments done in the article:
Tapio Pahikkala, Antti Airola, Sami Pietilä, Sushil Shakyawar, Agnieszka Szwajda, Jing Tang, and Tero Aittokallio. Toward more realistic drug-target interaction predictions. Briefings in Bioinformatics, 16(2):325-337, 2015. [Bib]
If you use the data and publish the results, please cite the above article as well as the original studies by Davis et al (2011) or Metz et al. (2011).
The below file contains the drug-target interaction affinities published by Davis et al (2011)
drug-target_interaction_affinities_Kd__Davis_et_al.2011.txt
The below files contain the drug-drug structural fingerprint similarities computed with the 2D and 3D Tanimoto coefficients using the Structure Clustering sever at PubChem, as well as the Extended-Connectivity Fingerprint similarities.
drug-drug_similarities_2D.txt
drug-drug_similarities_3D_SFS.txt
drug-drug_similarities_ECFP4.txt
The below files contain the target-target sequence similarities computed with the original and normalized versions of the Smith-Waterman (SW) score.
target-target_similarities_WS.txt
target-target_similarities_WS_normalized.txt
The rows (and columns) of the drug-drug similarity matrices correspond to the rows of the interaction affinity matrix and the rows (and columns) of the target-target similarity matrices correspond to the columns of the interaction affinity matrix.
The below files contain the PubChem CIDs of the drugs and gene names of the targets with the rows corresponding to those of the above files:
drug_PubChem_CIDs.txt
target_gene_names.txt
The below file contains the drug-target interaction affinities published by Metz et al. (2011). The file contains unknown entries marked with "nan".
known_drug-target_interaction_affinities_pKi__Metz_et_al.2011.txt
The below file contains the drug-drug structural fingerprint similarities computed with the 2D Tanimoto coefficients.
drug-drug_similarities_2D__Metz_et_al.2011.txt
The below file contains the target-target sequence similarities computed with the normalized Smith-Waterman (SW) score.
target-target_similarities_WS_normalized__Metz_et_al.2011.txt
In the interaction prediction results of the article, the target-target similarities were unfortunately mixed for the Davis et al (2011) data in the sense that the normalized WS similarities were used in place of unnormalized similarities, and vice versa. We thank Marten Heidemeyer for pointing this out and apologize for any inconvenience this may have caused.