This is a simple wrapper around R packages ontologyIndex and ontologySimilarity.
This script takes a table of genomic variants in some patients as an input and outputs 2 metrics for them:
- raw similarity between a gene and patient's phenotype (0 = no similarity, 1 = perfect match)
- phe-value - how well this particular gene explains the patient's phenotype, in comparison with other genes affected in this dataset (0 = this gene is the best match for this patient phenotype, 1 = this is absolutely random match and other genes explain the phenotype much better).
Requirements:
- both abovementioned R packages installed,
- genes_to_phenotype file from OMIM is located in the same folder as this script,
- your input file "affected genes in samples" is tab-separated, each affected gene-sample takes a separate row, and your file has a header with the column names
disease_details_HPO_term_id(here HPO terms OF PATIENTS should be described, separated by "; " split, for example - HP:0000248 - Brachycephaly; HP:0000343 - Long philtrum; , without quotation marks!),sample(containing sample IDs - take care that same IDs belong to patients with the same HPO described),gene(gene ID used in OMIM).
You run the script as:
Rscript match_genes_to_pheno.R input_file.txt
Result file with the name phevalues.annot.tsv will appear in the directory with the script.
The script will also produce a plot with phe-values and raw similarity scores - the general variants of interest are located on the top left part.