CriteriaForSuccess

The aggregation of Lichtarge Lab genotype-phenotype validation experiments

Installation

1. git clone https://github.com/kevwilhelm95/CriteriaForSuccess.git
2. conda env create -f environment.yml
3. conda activate pyCFS

Required arguments:

Argument	Description	Options
--ExperimentName	Name of disease and/or cohort	NA
--InputPath or --InputList	Path to BigPipeline result directory or single-column .txt file with genes of interest	NA
--PickExperiments	No-space comma-separated list of experiments to run	All (default), GS Overlap, nDiffusion, MGI, OR, PubMed Enrichment, Variants By Sample, InterMethod Connectivity, Pharmacology
--OutPutPath	Path to output directory	NA
--cores	Number of cores used to run	NA

Optional arguments:

Argument	Description	Options	Required for
--GSPath	Path to CSV of Gold Standard Lists	NA	All, GS Overlap, nDiffusion, InterMethod Connectivity
--VCF	Path to VCF	NA	All, OR
--PickNetwork	Network to use for nDiffusion/Intermethod Connectivity	STRINGv10, STRINGv11, MeTEOR (default: STRINGv11)	All, nDiffusion, InterMethod Connectivity
--PubMedKeywords	No-space comma-separated list of key words to query for co-mention (e.g. "Type 2 Diabetes,Insulin")	NA	All, PubMed Enrichment
--InterConnectivity_Evidences	No-space comma-separated list of evidence types	neighborhood, fusion, cooccurence, coexpression, experimental, database, textmining	All, InterMethod Connectivity
--InterConnectivity_Confidence	Edge weight to test for connections	all, medium, high, highest	All, InterMethod Connectivity
--CaseControlPath	Path to CSV; Col. 1- IDs, Col.2- 1,0 (Case,Control); No header	NA	All, OR, Variants By Sample
--AC_Threshold	Threshold of Pathways to incorporate into Consensus2 (Default = 5)	NA	Only used if --InputPath used
--ref	Genomic reference locations	GRCh37, GRCh38 (default), hg19, hg38	All, OR, Variants By Sample

Example Call

python CriteriaForSuccess.py
--ExperimentName Test_run
--InputPath /path/to/dir/
--PickExperiments All
--CaseControlPath /path/to/sample.csv
--OutPutPath /path/to/output/dir
--cores 1
--GSPath /path/to/GS.csv
--VCF /path/to/vcf.gz
--PickNetwork STRINGv11
--PubMedKeywords "Type 2 Diabetes,Insulin"
--InterConnectivity_Evidences "neighborhood,fusion,coexpression"
--InterConnectivity_Confidence highest
--CaseControlPath /path/to/samples.csv
--AC_Threshold 5
--ref GRCh38

Experiment Descriptions

Experiment	Description
GS Overlap	Number of Gold Standard genes recovered by experimental list. Significance tested by hypergeometric test
nDiffusion	Network connectivity against Gold Standard genes using nDiffusion algorithm (PMID: 31797167). Significance tested by Z score against 100 random, degree-matched gene sets
InterMethod Connectivity	First-neighbor enrichment to Gold Standard genes. Significance tested by Z score against 100 random, degree-matched gene sets
PubMed Enrichment	Co-mentions in title or abstract by query gene for defined keywords. Significance tested by Z score using 100 random gene sets
OR (Odds Ratio)	Allele frequency differences between cases/controls for by variants (Allele frequency > 1%) or by gene (Allele frequency < 1% variants collapsed by gene). Significance tested by Fisher's Exact Test
Variants By Sample	Parser to obtain .txt file containing information on which samples harbor which variants
MGI	Enrichment test for high-level phenotypic abnormalities observed in mouse-models obtained from the Mouse Genome Informatics resource. Significance tested by Fisher's Exact Test
Pharmacology	Pulls drug-protein interactions for query genes from DGIdb.org