initial summaries and scripts

Post date: Jan 01, 2016 9:54:35 PM

I have summarized and made plots for the first set of simulations (data sets 1-4, quantitative metric of fitness). I will however need to redo most of this with data sets 5-6 and the binary measures (which only include versions with functional variants in the data set).

Here are the key scripts, all in /pscratch/A01963476/sims/gemmasel/.

• 1. calpost.pl: calculates posterior summaries (median, 90% ETIPs, mode, 90% HDPIs) for PVE, PGE, and n_gamma. All results for one hyperparameter (i.e., all data sets and replicates) end up in one file (though I might use a separate one for binary). These are the pp_* files.

  2. summarizeSnpEffects.pl: performance diagnostics for SNPs (PIPs, correlations with true values, etc.). This is run on one data set at a time and generates the snp*txt files.

  3. plotPost.R: generate plots of point estimates and CIs for PVE and n_gemma and PGE by replicate for all data sets (four now, will want to add 5 and 6).

  4. summarizePost.R: calculate RMSE and 90% coverage for PVE and n_gamma for a table (will need to add 5 and 6, and binary); also makes a plot of the distribution of point estimates.

  5. plotSnpSummaries.R: creates violin plots of SNP performance metrics. New data sets (5 and 6) will need to be added.