StemCellCKB: An Integrated Stem Cell-Specific Chemogenomics KnowledgeBase for Target Identification and Systems-Pharmacology Research

Abstract

Given the capacity of self-renewal and multilineage differentiation, stem cells are promising sources for use in regenerative medicines as well as in the clinical treatment of certain hematological malignancies and degenerative diseases. Complex networks of cellular signaling pathways largely determine stem cell fate and function. Small molecules that modulate these pathways can provide important biological and pharmacological insights. However, it is still challenging to identify the specific protein targets of these compounds, to explore the changes in stem cell phenotypes induced by compound treatment and to ascertain compound mechanisms of action. To facilitate stem cell related small molecule study and provide a better understanding of the associated signaling pathways, we have constructed a comprehensive domain-specific chemogenomics resource, called StemCellCKB ( http://www.cbligand.org/StemCellCKB/ ). This new cloud-computing platform describes the chemical molecules, genes, proteins, and signaling pathways implicated in stem cell regulation. StemCellCKB is also implemented with web applications designed specifically to aid in the identification of stem cell relevant protein targets, including TargetHunter, a machine-learning algorithm for predicting small molecule targets based on molecular fingerprints, and HTDocking, a high-throughput docking module for target prediction and systems-pharmacology analyses. We have systematically tested StemCellCKB to verify data integrity. Target-prediction accuracy has also been validated against the reported known target/compound associations. This proof-of-concept example demonstrates that StemCellCKB can (1) accurately predict the macromolecular targets of existing stem cell modulators and (2) identify novel small molecules capable of probing stem cell signaling mechanisms, for use in systems-pharmacology studies. StemCellCKB facilitates the exploration and exchange of stem cell chemogenomics data among members of the broader research community.

Figure 1

StemCellCKB overview, including integrated computing and data-mining functions. The COMPOUND module includes two chemical records that describe small molecules specific to stem cell targets and small molecules involved in stem cell pathways. The TARGET module contains a list of proteins relevant to stem cell regulation and the HTDocking tool. The PATHWAY module describes signaling pathways relevant to stem cell regulation. The TOOLS module enables prediction of chemical properties, toxicity, and targets for small molecule stem cell modulators.

Figure 2

Interaction network of stem cells, pathways, targets, and compounds. Red rectangles represent stem cell types. Yellow nodes represent core signaling pathways involved in stem cell regulation. Green nodes represent core protein targets associated with stem cell modulation. Blue nodes represent known core stem cell-related compounds.

Figure 3

Chemogenomics data in StemCellCKB. (A) Association among compounds, protein targets, and signal pathways involved in stem cell regulation. (B) StemCellCKB target classification. (C) Compounds known to be involved in stem cell regulation were plotted according to stem cell type and the corresponding target pathways. (D) Compounds known to be involved in stem cell regulation were plotted according to their development phases and associated pathways. (E) Compounds known to be involved in stem cell regulation were plotted according to their protein targets.

Figure 4

StemCellCKB Database validation by target identification. (A) Chemical-structure input window of the HTDocking server, with CHIR99021 displayed. (B) HTDocking results. (C) Polypharmacology analysis of known stem cell drugs. Green nodes represent the known or biologically validated protein targets of the compounds, and pink nodes represent new targets predicted using HTDocking. (D) Interaction between Z48 (salmon stick) and GSK3β (rainbow cartoon) in the active site (PDB ID: 3I4B). (E) CHIR99021 (magenta stick) docked into the active site of GSK3β (rainbow cartoon; PDB ID: 3I4B).