GenMore Application

GenMore is a locally hosted application that allows users to upload their DNA SNP information to recieve ClinVar pathogenic match results. Users can then compare their disease risk associated markers with ancient individuals.

Resource Information

Original ClinVar data can be found at: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/\ variant_summary.txt.gz : working version was built with the 2026-03-10 downloaded version.
README: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/README

AADR data taken from the Reich Lab at: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/FFIDCW\ v54.1_1240K_public
The data was originally downloaded and converted to PLINK format by Eran Elhaik using the following:
#Get the data
wget https://reichdata.hms.harvard.edu/pub/datasets/amh_repo/curated_releases/V54/V54.1/SHARE/public.dir/v54.1_1240K_public.tar\ #Convertf
/home/eelhaik/Tools/EIG-5.0.2/bin/convertf -p /home/eelhaik/Tools/EIG-6.0.1/CONVERTF/par.EIGENSTRAT.PED.aDNA

Project Structure Information

Scripts Information

Pre-Filtering Scripts

1. clinvar_parser.py
    * Used to filter the variant_summary.txt file from ClinVar.
    * SNPs only
    * Pathogenic and Likely Pathogenic
    * Remove missing rsIDs
    * Keep only GRCh37 assembly (AADR dataset used GRCh37 assembly for SNPs)
    * Derived dominance information from phenotypic description

2. clinvar_check.sh
    * Used to check that the applied filters worked as expected.
    * Checks that all necessary columns are included.
    * Checks all variants are SNPs.
    * Checks only pathogenic and likely pathogenic SNPs are included.
    * Checks there are no missing rsID entries.
    * Checks for only Chr37 Assembly entries. 
    
3. AADR_parsing.py
    * Used to parse through the PLINK AADR files and match pathogenic ancient DNA SNPs with the parsed ClinVar data.
    * Ref allele cannot = alt allele.
    * Matches on rsID first then on alt allele.
    * Missing genotypes are dropped.
    * Carrier vs. affected info is ascertained from the phenotype list content (dominance is determined and disease status reflects zygosity at the SNP).
    * Modern individuals are excluded based on the included Modern_samples.txt individual names.
    * Non-exact alleles dropped (ex. R).

App Functionality Scripts

1. user_parser.py
    * Used to parse user input files into standardized input file format. 
    * Handles 7 different input file types.

2. clinvar_user_match.py
    * Used to match user SNPs to ClinVar filtered pathogenic SNPs.
    * Ref allele cannot = alt allele.
    * Matches on rsID first then on alt allele.
    * Carrier vs. affected info is ascertained from the phenotype list content (dominance is determined and disease status reflects zygosity at the SNP).

3. compare.py
    * Used to compare user output SNPs and ancient DNA AADR SNPs.
    * Load AADR_parsing output and clinvar_user_match output.
    * Merges the two datasets on rsid and individual_id.
    * Compares disease state determinations between the two datasets.
    * Outputs a TSV file with the comparison results.

4. app.py
    * Full script that calls on necessary functions for local app output.

Dependencies

Dependencies are included in the envs/ folder in "plink_parsing_env.yml" and "requirements.txt".
Notable packages and versions:

1. pandas-plink v.2.3.2
2. python v.3.13.11
3. streamlit v.1.55.0

Pre-parsed ClinVar dataset.

The ClinVar dataset was parsed prior to app running with the included script "clinvar_parser.py." The parsed output is included in the results/ folder.
The successful filtering was confirmed with clinvar_check.sh.

# call on python script for parsing
python scripts/clinvar_parser.py -f resources/variant_summary.txt -d results/

Pre-parsed AADR dataset.

The AADR dataset was parsed using pandas-plink and the python script "AADR_parsing.py".
The following files downloaded from the AADR database are needed to parse the ancient DNA.
The AADR original files were converted to PLINK friendly files so the alternate and reference allele order were swapped.
The parser handles this swap to ensure accurate alternate allele matching with ClinVar SNPs.

1. Modern_samples.txt - used to filter out modern individuals.
2. v54.1_1240K_public.bed - used for all SNP X individual information.
3. v54.1_1240K_public.bim - used to SNP and genotype information.
4. v54.1_1240K_public.fam - used for all individual ID index information.

# call on python script for parsing
python3 scripts/AADR_parsing.py -p resources/AADR/v54.1_1240K_public -c results/ClinVar_parsed.tsv -m resources/AADR/Modern_samples.txt -o AADR_clinvar_matches.tsv -d results/

Running GenMore.

After downloading this repository to your local computer, from the project root run the following command to open GenMore locally.
Ensure proper project structure as shown in the beginning of the README.
The pre-parsed ClinVar and AADR datasets as well as two sample inputs are included in the project directory so no
pre-filtering is needed as the user.

streamlit run app.py