deletions | additions       diff --git a/figures/rankplot_low458_high5000_T1D/caption.tex b/figures/rankplot_low458_high5000_T1D/caption.tex  index dcc5f23..f6e0292 100644  --- a/figures/rankplot_low458_high5000_T1D/caption.tex  +++ b/figures/rankplot_low458_high5000_T1D/caption.tex 
...
\label{fig:sig-thy-bootstrap-consistency}  Scatter plot indicating maximum rank over all bootstrap sub-samples vs. minimum rank in any bootstrap sub-sample for the bootstrap-consistent set of 458 SNPs (A), and a random sample of 5000 of the remaining SNPs (B). A total of 126519 SNPs (not included when generating these graphs) were unranked in at least one bootstrap sub-sample, as no genetic difference was observed between case and control groups with that SNP. The difference between minimum and maximum rank gives an indication of the reliability of a particular marker for association testing in a general population. Of those markers in the bootstrap-consistent set of 458 SNPs, 57\% were ranked in the top 5000 markers in all bootstraps. Of the remaining 489574 SNPs, 95\% (4464977) had a maximum rank of 350000 or more (including 126519 unranked SNPs).