Table 2.
Comparison of drug/target discovery tools.
| Algorithm/Input/Output | Approach | Main advantages | Limitations/biases/risks |
|---|---|---|---|
| PharmGKB [Barbarino et al. (25)] Input: list of genes or proteins Output: druggable genes |
An organised knowledgebase containing genetic, clinical, and cellular phenotype knowledge networks | 1. Data curated from PubMed, DrugBank, dbSNP 2. Accessible via API |
1. Does not have the facility to input a list of genes and generate pathways |
| DrugBank [Wishart et al. (26)] Input: single drug target Output: drug information |
Bio-/chemo-informatics resource that combines drug and drug target information | 1. Updated daily | 1. Cannot input a gene list or generate PPI networks |
| ChEMBL [Gaulton et al. (27)] Input: single drug target Output: drug information |
An open large-scale bioactivity database combining molecule, target, and drug data | 1. Provide web services for programmatic access to limited to specific queries | 1. Does not provide integrated PPI/drug/target outputs |
| TTD [Chen et al. (4)] Input: list of genes Output: related drug target and disease information |
Provides details on known therapeutic nucleic acid and protein targets, associated disease conditions, pathway information, and drug/ligand details | 1. Can search via target name, drug name, disease name and so on 2. Pathways use KEGG, MetaCyc, Reactome, Wikipathways |
1. Cannot create pathways via input gene lists |
| DisGeNET [piñero et al. (33)] Input: list of genes Output: summary/evidence of gene-disease associations |
Comprehensive archive of genes and variants associated to human diseases. Collection of data on genotype-phenotype relationships from several of the most popular resources in this area | 1. Database can be searched by target or drug/ligand names 2. Genes and variants associated to human diseases |
1. No PPI networks for input genes before drug target determination 2. Only focuses on genes/variants involved in human diseases |
| DGIdb [Cotto et al. (29)] Input: list of genes or proteins Output: associated drugs |
A collection of drug–gene interactions and gene druggability information | 1. Combines drug–gene interactions and possible druggable genes 2. Combines data from NCI, PharmGKB, TTD, GO, DrugBank etc 3. Can upload gene list |
1.Unable to generate the PPI network 2.Does not capture disease data |
| PHAROS [Nguyen et al. (31)] Input: interested list of genes or proteins Output: gene ontology terms, pathways, drugs, diseased grouped by the input gene |
Is a web interface which presents data from the Target Central Resource Database (TCRD) which collates many heterogeneous gene/protein datasets | 1. Considers all human protein targets 2. Information on understudied targets 3. Offer programmatic access to the data 4. Consists of protein expression data, disease and phenotype associations, bioactivity data, drug target interactions 5. Includes GO terms |
1. Unable to upload gene lists 2. Data output is not grouped based on GO terms (thus does not enable rapid assessment of PPI networks and associated drugs based on key biological processes or molecular functions performed by the cell of interest) 3. Human-specific |
| GPSnet [Cheng et al. (32)] Input: list of cancer genes Output: drug target network |
Network based, integrated algorithm for cancer related diseases and approved or investigational drugs | 1. Drug-target networks 2. Uses GSEA algorithm in gene enrichment |
1. Focus only on cancer-specific diseases 2. No free API access |
Comparison of drug/target discovery tools (most developed pipelines shown in grey).