Table 1 Comparison of PCAN to related methods
From: PCAN: phenotype consensus analysis to support disease-gene association
Software | Application | Approach | Description | Availability |
|---|---|---|---|---|
PCAN | Gene- phenotype exploration | Indirect, phenotype-based | Implements a readily interpreted, statistical definition of phenotype consensus for configurable lists of mechanistically-related genes. Can be used for gene-prioritisation and also versatile, trait-level exploration of gene-phenotype relationships within pathways and biological networks. | R package |
CSI-OMIM [11] | Disease diagnosis | Direct, phenotype-based | Improved phenotype searching of NLP processed OMIM Clinical Synopsis descriptions. Phrases are tagged with ontological terms (MESH, UMLS) and clustered into groups of synonymous expressions. | Website |
Phenomizer [10] | Disease diagnosis | Direct, phenotype-based | Improved phenotype searching using semantic similarity methods based on HPO annotations for rare diseases. | Website |
PhenoDigm [29] | Disease-gene prioritisation | Direct, phenotype-based | Gene prioritisation based on phenotype comparison across model organisms. Model organism trait ontologies (e.g. HPO and MPO) are cross-linked and semantic similarity is computed using the OWLSim algorithm. | Website |
Exomewalker [12] | Disease-gene prioritisation | Indirect, phenotype-based | Performs a random walk of the STRING protein-interaction network, seeded with genes linked to diseases with a high semantic similarity to the disorder under investigation. Genes are prioritised based on the random walk score and variant-level criteria combined using a linear model. | Website and command line (via Exomiser) |
Syndrome to Gene [32] | Disease-gene prioritisation | Indirect, ontology-based | Use CSI-OMIM to identify genes that cause similar diseases. Quantify gene-relatedness by comparing information vectors derived from 18 source databases using a Jaccard similarity coefficient. Genes are prioritised if they are related to genes that cause similar phenotypes. | Website |
OVA [13] | Variant prioritisation | Indirect, ontology-based | Generates extensive, gene-level, multi-ontology annotation profiles for candidate variants and a query phenotype. Direct gene annotations are supplemented with inferred annotations from model organism orthologues and network neighbours. Annotation vectors are compared by computing domain-specific semantic similarities and combined using a Random Forest model to rank variants. | Website |
Exomiser [14] | Variant prioritisation | Pipeline | Variant ranking is based on both variant-level properties (allele frequency, pathogenicity) and gene-level semantic similarities for directly linked human diseases, model organism phenotypes as well as network proximity to similar phenotypes using ExomeWalker. | Website and command line |