Abstract
Background: Dermatofibrosarcoma protuberans (DFSP) is a rare mesenchymal tumor that is primarily treated with surgery. Targeted therapy is a promising approach to help reduce the high rate of recurrence. This study aims to identify the potential target genes and explore the candidate drugs acting on them effectively with computational methods.
Methods: Identification of genes associated with DFSP was conducted using the text mining tool pubmed2ensembl. Further gene screening was carried out by conducting Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. Protein-Protein Interaction (PPI) network was constructed by using the Search Tools for the Retrieval of Interacting (STRING) database and visualized in Cytoscape. The gene candidates were identified after a literature review. Drugs targeting these genes were selected from Pharmaprojects. The binding affinity scores of Drug-Target Interaction (DTI) were predicted by a deep learning algorithm Deep Purpose.
Results: A total of 121 genes were found to be associated with DFSP by text mining. The top 3 statistically functionally enriched pathways of GO and KEGG analysis included 36 genes, and 18 hub genes were further screened out by constructing a PPI networking and literature retrieval. A total of 42 candidate drugs targeted at hub genes were found by Pharmaprojects under our restrictions. Finally, 10 drugs with top affinity scores were predicted by DeepPurpose, including 3 platelet-derived growth factor receptor beta kinase (PDGFRB) inhibitors, 2 platelet-derived growth factor receptor alpha kinase (PDGFRA) inhibitors, 2 Erb-B2 receptor tyrosine kinase 2 (ErbB-2) inhibitors, 1 tumor protein p53 (TP53) stimulant, 1 vascular endothelial growth factor receptor (VEGFR) antagonist, and 1 prostaglandin-endoperoxide synthase 2 (PTGS2) inhibitor.
Conclusion: Text mining and bioinformatics are useful methods for gene identification in drug discovery. DeepPurpose is an efficient and operative deep learning tool for predicting the DTI and selecting the drug candidates.
Keywords: Dermatofibrosarcoma protuberans, drug discovery, text mining, deep learning, deep purpose, DTI.
Graphical Abstract
[http://dx.doi.org/10.1016/j.jaad.2021.08.010]
[http://dx.doi.org/10.1111/bjd.19346]
[http://dx.doi.org/10.1016/j.jaad.2012.01.039]
[http://dx.doi.org/10.1111/1346-8138.15056]
[http://dx.doi.org/10.1016/j.jaad.2021.06.856]
[http://dx.doi.org/10.1001/jamanetworkopen.2019.10413]
[http://dx.doi.org/10.1016/j.jaad.2012.10.040]
[http://dx.doi.org/10.1016/j.jaad.2018.05.016]
[http://dx.doi.org/10.1001/jamadermatol.2018.4940]
[http://dx.doi.org/10.1158/1078-0432.CCR-13-1411]
[http://dx.doi.org/10.1016/j.ctrv.2021.102259]
[http://dx.doi.org/10.1111/bjd.19837] [PMID: 33835484]
[http://dx.doi.org/10.1177/1060028017702406] [PMID: 28707550]
[http://dx.doi.org/10.1016/j.abb.2020.108730]
[http://dx.doi.org/10.1021/acs.accounts.7b00083]
[http://dx.doi.org/10.1038/nrg3337] [PMID: 23150036]
[http://dx.doi.org/10.1038/nbt.4233]
[http://dx.doi.org/10.1038/s41573-020-00117-w]
[http://dx.doi.org/10.1093/bioinformatics/bty593] [PMID: 30423097]
[http://dx.doi.org/10.1093/bioinformatics/btaa1005]
[http://dx.doi.org/10.1007/s00018-004-4272-z] [PMID: 15583853]
[http://dx.doi.org/10.1016/j.phrs.2018.01.021]
[http://dx.doi.org/10.1001/jamadermatol.2015.2389]
[http://dx.doi.org/10.1038/modpathol.3800783]
[http://dx.doi.org/10.1111/j.1365-2133.2006.07146.x]
[http://dx.doi.org/10.1097/PAS.0b013e3181cf7fc1] [PMID: 20139758]
[http://dx.doi.org/10.1016/j.humpath.2017.07.001]
[http://dx.doi.org/10.3390/cancers13092224]
[http://dx.doi.org/10.1097/00000478-200604000-00002] [PMID: 16625088]
[http://dx.doi.org/10.1186/s13287-021-02512-5]
[http://dx.doi.org/10.1111/j.1600-0560.2007.00855.x]