I'm working on a project using next-gen sequencing to fine-map a genetic association in a gene region. Now that I've sequenced the region in a small sample, I'm picking SNPs to genotype in a larger sample. When designing the genotyping assay the lab will need flanking sequence. This is easy to get for SNPs in dbSNP, but what about novel SNPs?

Lately I haven't written as many full length posts as usual, but here's a quick roundup of a few links I've shared on Twitter (@genetics_blog) over the last week: First, 23andMe is having a big DNA Day Sale ($108) for the kit + 1 year of their personal genome subscription service https://www.23andme.com/. Previously mentioned R IDE RStudio released a new beta (version 0.93) that includes several new features and bugfixes.

Hansong Wang, our biostats professor here at the Hawaii Cancer Center, generously gave me some R code that goes through a SNP annotation file (i.e. a mapfile) and selects SNPs that are at least a certain specified distance apart. You might want to do this if you're picking a subset of SNPs for PCA, for instance.

I recently started using RStudio, the amazing new IDE for R. You can view all of RStudio's keyboard shortcuts by going to the help menu, but I made this printable reference for myself and thought I'd share it. I only included the Windows shortcuts, and I cut out all the obvious ones (Ctrl-S for save, Ctrl-O for open, etc) so it would fit neatly on one page.

The *ABEL suite of R packages and software for genetic analysis has grown substantially since the appearance of GenABEL and the previously mentioned ProbABEL R packages. There are now a handful of useful R packages and other software utilities facilitating genome-wide association studies, analysis of imputed data, meta-analysis, efficient data storage, prediction, parallelization, and mixed model methods.

Abhijit over at Stat Bandit posted some nice code for making forest plots using ggplot2 in R. You see these lots of times in meta-analyses, or as seen in the BioVU demonstration paper. The idea is simple - on the x-axis you have the odds ratio (or whatever stat you want to show), and each line is a different study, gene, SNP, phenotype, etc.

In my previous post I showed you how to randomly split up a dataset into training and testing datasets. (Thanks to all those who emailed me or left comments letting me know that this could be done using other means.

Just saw the announcement of the availability of Rstudio, a new (free & open source) integrated development environment for R that works on Windows, Mac, and Linux. Judging from the screenshots, it looks like Rstudio supports syntax highlighting for Sweave &

I recently analyzed some data trying to find a model that would explain body fat distribution as predicted by several blood biomarkers. I had more predictors than samples (p>n), and I didn't have a clue which variables, interactions, or quadratic terms made biological sense to put into a model.

When I have a question I usually ask the internet before bugging my neighbor.