Pharos 2023: an integrated resource for the understudied human proteome

Abstract

The Illuminating the Druggable Genome (IDG) project aims to improve our understanding of understudied proteins and our ability to study them in the context of disease biology by perturbing them with small molecules, biologics, or other therapeutic modalities. Two main products from the IDG effort are the Target Central Resource Database (TCRD) (http://juniper.health.unm.edu/tcrd/), which curates and aggregates information, and Pharos (https://pharos.nih.gov/), a web interface for fusers to extract and visualize data from TCRD. Since the 2021 release, TCRD/Pharos has focused on developing visualization and analysis tools that help reveal higher-level patterns in the underlying data. The current iterations of TCRD and Pharos enable users to perform enrichment calculations based on subsets of targets, diseases, or ligands and to create interactive heat maps and UpSet charts of many types of annotations. Using several examples, we show how to address disease biology and drug discovery questions through enrichment calculations and UpSet charts.

Figure 1.

TDL changes over time • Chart of the TDL changes between an early version (TCRD v3.0), a version at the time of the last update publication (TCRD v6.7), and the current version (TCRD 6.13). The decrease of Tdark and subsequent increase of other development levels shows an overall increase in target illumination.

Figure 2

Target details updates. (A) Expression data from five sources is color coded according to expression type (red: aggregate score, blue: RNA expression, yellow: protein expression) and displayed as a heat map on the left panel. Cells in the heat map display details about the expression data for each tissue when clicked. The right panel contains tabs to show either a circular treemap, where tissues are grouped according to the UBERON hierarchy, and an anatomogram, where tissues on a human form are shaded according to their expression level. The circular treemap is interactive and can be used to filter the heat map. (B) Text-mined target-disease associations from TIN-X are shown in an interactive scatter plot of importance vs novelty adjacent to a circular treemap. The circular treemap groups the associations based on the hierarchy defined by the Mondo Disease Ontology. Selecting a circle in the right panel highlights corresponding points on the scatter plot for those diseases. This dynamic visualization helps users find classes of diseases which tend to be both high in the importance and novelty metrics. (C) GWAS traits associated with a single target, scored and ranked according to the TIGA data processing pipeline. More reliable associations tend to have a higher Mean Rank Score and a higher Beta Count.

Figure 3.

Structure search. The component contains a Marvin JS widget showing an editable query structure. Structures can also be edited via the Query SMILES input or resolved using an external tool (not pictured). This serves as a starting point for performing a similarity, or substructure, search of compounds in TCRD, or a search for predicted targets, using the QSAR models in NCATS Predictor.

Figure 4.

Enrichment analysis. (A) Left shows the number of targets associated with different diseases in a full target list. Right shows the number of targets associated with different diseases in a target list consisting of targets with a documented protein-protein interaction with DRD2. Note how there is a lot of overlap between the entries in this list, when the list is sorted by the naive counts. (B) Enrichment score results after performing Fisher's Exact Test on the filtered list. Note how the top entries in the list comprises a different set of diseases, including a lot of neurological disorders, and substance dependence disorders, diseases which are more commonly known to be related to DRD2.

Figure 5.

UpSet chart examples. (A) For a list of ligands, this plot shows different combinations of targets against which each ligand has been shown to be active. (B) For a target list, this plot shows how many targets in the list were on each combination of NIH lists. Gold highlighting shows filters that are currently applied to the list, so this list would consist of 12 targets that were on the IDG list from 2017 to 2020, and removed in 2022. (C) In this case, the corresponding target list would show 385 targets that have data from ProKinO, but not from Dark Kinase Knowledgebase. (D) In this case, the corresponding target list would show 21 targets annotated by PANTHER to be a RNA binding protein, but not a DNA binding protein.