search

BridgesLab / CushingAcromegalyStudy

The source code for the cushing and acromegaly studies, currently ongoing
Other
1 stars 2 forks source link

Retrospective power analysis #62

Closed davebridges closed 9 years ago

davebridges commented 9 years ago

For the response letter, Id like to state something along the lines of "based on our design we would have been able to assign significance to XXX% of genes that were altered by 50% or more (q=0.05)" Basically a sense of the power (without adjusting for age etc). Could you do a quick calculation. I want to do this as part of the argument that its worthwhile to mention genes that weren't quite significant.

qtran1 commented 9 years ago

Ok! I can do that!

Quynh T. Tran, PhD

On Jun 21, 2015, at 9:57 AM, Dave Bridges notifications@github.com<mailto:notifications@github.com> wrote:

For the response letter, Id like to state something along the lines of "based on our design we would have been able to assign significance to XXX% of genes that were altered by 50% or more (q=0.05)" Basically a sense of the power (without adjusting for age etc). Could you do a quick calculation. I want to do this as part of the argument that its worthwhile to mention genes that weren't quite significant.

— Reply to this email directly or view it on GitHubhttps://github.com/BridgesLab/CushingAcromegalyStudy/issues/62.

qtran1 commented 9 years ago

We have 23% power in detecting the target effect size of 1.5 (50% change in expression). The arguments are as follow:

  1. average depth of coverage is 10
  2. sample size in group 1 is 11 (control)
  3. sample size in group 2 is 5 (cushing's)
  4. biological coefficient of variant in control is 0.60
  5. biological coefficient of variant in Cushing's is 0.49
  6. alpha = 0.05

These information can be found in the power-analysis script.