Project Member

Blake Haas

The goal of this quarter

Determine the optimal multi-stage design to maximize power for a given cost.

The Schedule for this quarter

week 1: Determine what math needs to be done in order to solve this problem (no programming)

week 2: Program to receive all needed variables from user, start writing functions I will call in program, find out how to code a simulation

week 3: Write more functions, determine when to use these functions

week4: finish program

Project Description

I will take in information from the user about the multistage study, and determine the optimal cost-effective way to maximize power. The optimal cost-effective study is a study that maintains almost all the power (~97%) of a 1 stage study while minimizing cost.

Related Papers

http://www.ncbi.nlm.nih.gov/pubmed/17549752?ordinalpos=1&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum

April 24

1) I made this page this week

2) Next week I will determine what math needs to be done in order to solve this problem (no programming)

3) It went according to plan this week

4) I deserve an A since I'm totally awesome.

May 1

1) I determined how I will code my program this week. The user will input total # of subjects, cost per SNP for a whole genome SNP chip, the cost per SNP for a specially designed SNP chip, and the threshold for determining whether or not to keep a SNP. The program will return Power vs. #SNPs genotyped and the cost vs. # of SNPs genotyped for each phase of the study. The program will also return the optimal # of SNPs to genotype for each phase of the 3 part study, which will be determined by maximizing the Power/Cost ratio.

2) Next week I will begin coding the program

3) It went according to plan this week

4) I deserve an A because I finished what I planned to accomplish this week.

May 9

1) I determined how I will calculate all the math that will be used in my program.

2) Next week I will begin coding the program

3) It didn't go according to plan, since I had an extremely busy week in lab so I didn't get to start programming. I will do more work next week.

4) I deserve an A- since I did as much as I could with the given time that I had this week.

May 16

1) I determined to change how i will do the project. I will calculate the power by observing the power after a user inputted number of simulations for a 2 stage (or 3 stage with time) study.

2) Next week I will code the program

3) It did not go according to plan since I did not plan to redo the project in a different way

4) I deserve an A for effort since I am now on the correct track and know what i will be doing.

May 23

1) I am almost done computing optimal power for multiple stages.

2) next week I will calculate optimal power based on simulations

3) It went according to plan this week

4) I deserve an A because I am putting in a lot of time.

May 30

1) I am still working on optimal power calculations for 2 and 3 stage studies. I will loop through all combinations of beta values and people in each stage to determine the optimal study design to maintain power (~97% of a 1 stage study) while keeping the cost the absolute lowest possible. I decided not to do the simulation portion because of time constraints.

2) Next week I will finish debugging and present it to the class

3) It went according to plan this week.

4) I deserve an A since I am working hard and almost done with the project.