Manuscript revisions

[pbreheny]

• [x] "directly jumps into a statistical question" More discussion of application in general, and especially up front.
• [x] "should be clearer descriptions on when the reviewed approaches can/cannot be applied"; be more explicit here?
• [x] More background on lasso
• [x] More discussion of software. Add a small section?
• [x] More discussion of nonlinear models. In particular, is logistic regression practical?
• [x] "may be of interest to know more about the statistical properties"...I don't know what this means. Like, theoretical properties? Maybe we can discuss? Mention Rohe and/or Buhlmann?
• [x] "Data examples are recommended"; "more realistic simulation scenarios...100,000s of SNPs"; Let's handle this with a genome-scale single replication study to address both points. Simulate vanderwas phenotype + spike in 50,000 snps with effect on simulated phenotype (select 1000)? How well do these methods do at finding these SNPs? Like a real data analysis... case study - single replicate of a genome wide analysis here's how it turned out. Compare with regular lasso, PC-lasso.
• [x] Finalize figures, check PRC curve issues for PC-lasso
• [x] Write methods section for genome-wide analysis
• [x] Write results section for genome-wide analysis
• [x] penalty parameter selection; also a tough one. We need to say more, but not a great deal is known.
• [x] Fig1, Table1 captions should specify # of replications and which SNPs are included (all 1000 or only those selected?) - do this for all figures, not just Fig1 Table1?
• [x] "For Figure 1, it would be helpful to have error bars included in bar plots." - described in updated caption. additional line in legend seems redundant.
• [x] "there is a typo in Eqn 15."
• [x] Get latex version working on genetic epi portal
• [x] Patrick to clip pdf tikz plots and convert to eps files
• [x] Point-by-point responses
• [x] New cover letter
• [x] Resubmit

[areisett]

Going to close this since we finished everything and have published