Please note that enrichment testing is an optional step in MendelVar, which runs only if any of the ontologies is selected from the list. Foregoing this step returns just the disease_overlap.txt and variant_overlap.txt tables. Disease_overlap.txt lists all the Human Phenotype and Disease Ontology (HPO, DO) terms associated with a given Mendelian condition overlapping the target genomic interval. ****
MendelVar enrichment testing options
<aside> 💡 INRICH takes a list of associated genomic intervals and tests for enrichment against gene sets. The test genomic intervals need not be independent from each other, as overlapping intervals are merged, as well as overlapping genes belonging to the same gene set. Background ('null') interval sets for enrichment testing are picked to match the test set in terms of the number of SNPs, SNP density and number of overlapping genes.
</aside>
MendelVar uses positions of biallelic 1000 Genomes variants with unique positions (similar to variants used in LDLink) in one of the six target populations: CEU (Utah residents of northern European descent), EUR (European), EAS (East Asian), SAS (South Asian), AFR (African) or AMR (Ad-mixed American**)** as background for INRICH SNP map file which is used for finding matching background intervals for enrichment testing.
INRICH can be run in two modes: gene and interval. In the gene mode, the enrichment statistic is calculated over the number of genes in any interval overlapping a given ontology term. In the interval mode, the enrichment statistic is calculated over the number of intervals with at least one gene, regardless of gene number, overlapping a given ontology term. Thus, INRICH in gene mode will upweight contribution from intervals that have more than one gene matching the same ontology term. In general, gene mode should result in higher power to detect enrichment in the context of MendelVar use case, if we expect multiple genes with related functions to cluster in a single locus. However, in the case of post-GWAS annotation analysis, any enrichment statistic calculated that way will be biased by double counting of the same GWAS signal.
MendelVar runs INRICH with the following settings:
inrich -c -r 50000 -q 10000 -p 1
where:
-c disables a memory intensive and bug-prone pre-computation step
-r specifies number of permutations for calculating empirical p-values
-q specifies number of permutations for correcting empirical p-values for multiple testing
-p specifies maximum p-value results to be printed to output