Team 2

Streamlining VCF data flow

Members:

• Thibaut Jombart (team lead)
• Emmanuel Paradis
• Klaus Schliep
• Jerome Goudet

Goals

• goals go here

Status

Day 2:

• Reviewed what is available in R for working with VCF files. There is one package (popgenome), not very easy to use.
• Feedback sought: how big will your data files be in 5 years from now?
  • my guess is 10⁶ loci, on hundreds to thousands individuals (we are already at 10⁵ loci, and sequencing costs keep falling) [jerome]
• Plan next to optimize geneind code to reduce memory consumption.
• Plan to interface hierfstat with adegenetand pegas by making use of the class geneind and loci

Day 3:

• able to read VCF for 1000 Genome project in less than a minute (not including genotype).
• reworking package '5' to load data faster, including ploidy
• now looking into how data can be moved faster and more seamlessly into hierfstat from adegenet
• Can perhaps also look into being compatible for individual-lacking VCF files? (Such as those from 1001 Arabidopsis genomes)

Day 4:

• Finalizing fast scanning and reading of VCF files. 1M loci in just a few seconds. Cleaning up code.
• Simplified data structure in adegenet. May break some code that depended on earlier versions. Need help for testing. If you find problems, please file issue on Github.
• Added function for genetic distances in hierfstat. Also discovered numeric type bug that is being fixed now. hierfstat is on Github now.

Products

New R packages:

• apex: Extension of the R package ape for multiple genes

Updates to R packages:

• adegenet: R Package for the Multivariate Analysis of Genetic Markers
• hierfstat: Estimation and tests of hierarchical F-statistics
• pegas: Population and Evolutionary Genetics Analysis System