A library for genome-wide associations using k-mers

A library for running k-mers-based GWAS

Binaries for beta version can be downloaded from here.

Details on how to create a k-mers table and how to run GWA can be found in the manual.

key functionalities

Run pipeline:

• python2.7 src/py/pipeline.py - Running the k-mers-GWAS pipeline

Build the k-mers table:

• kmers_add_strand_information - Create k-mers list from KMC output
• list_kmers_found_in_multiple_samples - Create a list of k-mers to use in the k-mers table
• build_kmers_table - Build the k-mers table

Other functionalities:

• kmers_table_to_bed - Convert the k-mers table to PLINK binary format
• filter_kmers - Textual output the presence/absence patterns of a set of k-mers
• emma_kinship_kmers - Calculate the kinship matrix from PLINK binary format
• emma_kinship - Calcualte the kinship matrix of the k-mers table

Prerequisites:

• Linux system with a 64 bit CPU
• R
• python2.7
• KMC (part of the release under external_programs/ directory)
• GEMMA (part of the release under external_programs/ directory)
• R packages (if not present the pipeline will try to automatically install):
  • MASS
  • Mvnpermute
  • matrixcalc

Examples:

In the examples directory there are two examples of how to use the library:

1. Pre-existing k-mers table, only run the k-mers-base GWA - Using the k-mers table for A. thaliana 1001G to run k-mers-based GWA on flowering time (same as Fig. 1 in our paper). This table can be used also to run GWA on other phenotypes measured on accessions part of the 1001G.
2. Building the k-mers table and using it to run GWAS - Building the k-mers table on 241 E. coli accessions (from Earle et al. 2016) and then running the GWA using this table.

Citing

If you used our library for published work, please cite us:

Identifying genetic variants underlying phenotypic variation in plants without complete genomes

Yoav Voichek and Detlef Weigel (2019).