canSAR: update to the cancer translational research and drug discovery knowledgebase.

The Cancer Target Association highlights page for EGFR, taken from canSAR. (A) Gene expression data comparing EGFR in normal non-cancer samples and the different AJCC pathological stages of lung cancer, showing a gradual increase in EGFR gene expression with progression of the disease; (B) Pie chart shows copy number aberrations (CNA) in the TCGA LUAD study, followed by bar charts showing the distribution of different CNAs along pathological stages. The drop-down menu allows users to navigate between studies and cohorts; (C) Lollipop plot showing mutation frequency in lung cancer patients along the EGFR amino acid sequence. Users can zoom into a segment of the display to explore this region of the sequence for evidence of functional role. Tracks under the sequence correspond to functional domains, post-translational modifications, ligand/drug binding, protein-interface hotspots, ligandable cavities etc. This interactive tool was developed to enable researchers to generate testable hypotheses for the role of a mutation of interest.

Data supporting technical feasibility assessment for EGFR, taken from canSAR. Analysis of technical feasibility can be explored in terms of target tractability and experimental practicability (Supplementary Figure S2). (A) Available drugs – as this is a clinically-precedented drug target, approved drugs and investigational drugs are shown at the top. Where canSAR is able to recommended chemical probes, these are also shown with links to the primary chemical probe resources (Chemical Probes Portal and Probe Miner); (B) A summary of the structure-based ligandability. The protein sequence is represented along the x-axis and known functional domains shown on top. A green histogram shows where predicted ligandable cavity positions lie; in this case, illustrating their clustering on the catalytic kinase domain; further details and interactive exploration of protein 3D structures and ligandable cavities can be reached through the ‘explore more’ button; (C) Summary of the orthogonal ligandability/target likeness assessments calculated in canSAR. Dials represent the percentile ranking of EGFR in each assessment: network-based assessment determines how closely the target resembles drug targets from oncology or other disease areas; ligand-based assessment assesses a target based on the properties, number and bioactivity of its ligands and bioactive molecules; precedence-based assessment explores family similarity to known drug targets; and finally an indication of whether the target is available to extracellular antibody or other biotherapeutics. These methodologies utilise predictive machine learning algorithms and are calculated for most proteins in the human proteome, thus enabling the assessment of novel, potentially druggable targets.

Segments of the new compound synopsis pages, illustrated here for the anti-hormonal breast cancer drug tamoxifen. (A) Chemical compound data showing structural correspondence. First, the relationship between the dosed ingredient which is a pro-drug, and the active form of the compound, the main, 4-hydroxy metabolite afimoxifene. Then, the parent compound, stripped of stereochemistry, allows grouping of different enantiomers as well as different salt forms of the compound. The freebase canonical tautomer layer shows three distinct canonical tautomers for all forms of tamoxifen in the group compounds that are chemically equivalent but represented differently in source databases. Moreover, the structure allows the identification of related enantiomers and molecules with undefined stereochemistry. In this case, our pipeline not only identifies R/S stereochemistry but also E/Z stereochemistry, with the three E/Z/undefined alkene stereoisomers shown. While the bioactivity data for these should not be mixed, the display alerts the user to related compounds that may hold useful information; (B) Clinical trials information. Users can navigate increasing details of compound information. For example, users can instantly link to clinical trials for the compounds where they exist. An interactive clinical trials view allows users to navigate and filter clinical trials based on disease, phase and other trial characteristics.