Snappy: A flexible SNP proxy finder

Features

• Find SNP proxies using multiple methods: exact match, position lookup, and linkage disequilibrium.
• Updates dbSNP identifiers.
• Compatible with Unix/Linux pipe operator (|).

Limitations

• Does not resolve duplicate dbSNP identifiers, such as those on the sex chromosomes.
• Analysis is restricted to autosomal chromosomes.
• Default human genome build is hg19 (can be modified in source).

Requirements

• Unix/Linux tool chain: grep, awk,bash, etc.
• sqlite3.
• curl & wget.
• perl.
• plink and genotype reference panel in PLINK binary format, one file per chromosome ([1-22].(bed|bim|fam).

Installation

1. Make executable, chmod +x snappy.
2. Copy snappy script to program directory that is included in $PATH, eg. /usr/local/bin or ~/bin/.
3. Build dbSNP database: snappy build_dbsnp. This step will take up to 3-5 hours to complete.

NOTE: If you move snappy to another location, move snp150.db (generated in step 3) to the same location.

Usage

Snappy takes as it's first argument one of the following subcommands:

• match: Matches rsids by exact lookup.
• update: Updates unmatched rsids to latest rsids from dbSNP.
• position: Match remaining unmatched rsids in query to SNPs in target that are reported by position (eg 1_123456.* or 1:123456.*).
• proxy: Match remaining umatched rsids to target rdis using LD to 1000 Genomes EUR, selecting the highest r2, if multiple, then the nearest.

Snappy can pass output to itself using unix pipe (|), so you can run one or more of the above analyses in the order you prefer.

Here is an example usage scenario:

snappy init query.txt target.txt | snappy match | snappy update | snappy match | snappy position | snappy proxy  > snappy.out

In the above snappy does the following:

1. Find all exact matches.
2. Updates remaining rsids using dbSNP.
3. Find all exact matches in remaining SNPs.
4. Lookup remaining SNPs by position.
5. Lookup remaining SNPs by proxy.

The first step is always snappy init query.txt target.txt, where query.txt is a file with rsids, usually lead variants from a GWAS (example MR exposure GWAS).

The target.txt file is a 2nd list of rsids, usually the complete list of rsids from a 2nd GWAS (example MR outcome GWAS).

Example output:

QueryID     TargetId         r2        Dist   MatchType
rs10030035  rs10030035       1         0      match
rs10030217  rs10030217       1         0      match
rs10033900  rs10033900       1         0      match
rs10038416  5_147593231_A_G  1         0      position
rs1005283   rs1005283        1         0      match
rs1007914   rs2093989        0.979852  12322  proxy