Genes for Good

This is a working page devoted to the Genes for Good project. It contains a description of the study and many working documents.

Genes for Good is an attempt to create a research design that can scale to tens of thousands or even millions of participants with minimal cost. We will ask each participant to provide basic information about themselves, such as their name and address, and will send them a spit kit. We'll type their saliva for exome+gwas markers and ask them to fill out a variety of surveys about themselves.

At the outset we'll use an app within Facebook to administer questionnaires, allowing us to take advantage of Facebook's existing social network to snowball sample. While there are lots of possibilities for data collection, we'll focus primarily on health, anthropometrics, behavior, and environment using questionnaires.

Meeting Minutes

July 25 2014

ACTION ITEMS

• Kevin

1. We need a warning for individuals who are eligible for a spit kit but who have not entered their address. Also an option to not receive a spit kit and therefore not receive warnings about the address ---- Kevin need to figure out where to trigger that check. Otherwise done.
2. Create web interface and sql database (or additional tables in existing database) to track spit kits.
3. Create a way for participants to download their phenotype data.
4. Create a way for participants to download their genotype data (the dummy file is here: /net/plummer/dummy.vcf.gz, which is a bgzipped text file)

• Scott

1. Check with Goncalo about the spit kit and the chip.

• Kate

1. Work on implementing Spatial N-back (or possibly category switching).
2. Implement the verbal IQ test.

• Jinxi

1. Decide whether you want to do a "Dashboard" panel of plots like in the bam.iobio.io site or to simply create a handful of new plots for each questionnaire.
2. Finish ancestry plot by putting participant's value on the plot.

• Johanna

1. Storage:looks like Dan will print labels for Bob, Johanna will send Dan the barcode numbers
2. Work with Kevin on the spit kit database
3. Work with Greg Zajac on developing pipeline to move and QC the genotype data from the sequencing core: Sean needs 2 weeks to set up storage space for us, and he needs to know how much space we want.

• Matt

1. Ancestry decomposition that splits different ancestral haplotypes into their ancestral groups.
2. Getting ancestry proportions from STRUCTURE, as Sebastian said.

• Some other ideas that came up that we should continue to consider, but will require additional work.

1. Tell a person (e.g., on the "dashboard") that their data has been used in X studies, and perhaps which studies/publications used their data.
2. allow participants to query their genome, and then annotate the result from UCSC api. Also provide them with summary information about their genome (educational aspect of app).
3. If we do IQ, say it's in beta mode and don't provide results.
4. Should age bins in the profile be smaller?
5. the 60-70 age bin should read 61-70.
6. A tool that statistically tests whether two health indicates are correlated over time.

• Next meeting: We'll set up a teleconference for Wed. August 13.

Ideas for Plots

CAVEAT "To protect privacy for those that first join the study, we will not compare participants to others before we have 250 participants. We will also exclude the 4 most extreme individuals from each graph."

1. To represent number of complete surveys over time. Fun, not totally useful, but that's not important for surveys. http://bl.ocks.org/mbostock/1256572
2. Track how participants move through the app: http://bl.ocks.org/kerryrodden/7090426
3. If we had high usage we might use this to track app activity in real time (actions, users logged in, etc.): http://square.github.io/cubism/

Database Design

The database design diagram, data field definition is available at https://docs.google.com/a/umich.edu/file/d/0B22ZPHMStlJjMk9PV2UxMWV4TGM/edit.

Study Design

Click on the images here to see flowcharts diagramming the study design.

Flow Chart

Motivating Participants

How do we motivate and reward participants? Here's one useful review from the Cochrane group. that lists the following (Jennifer)

1. Electronic surveys (based on 32 eligible trials)
   1. Give them information back (OR=1.36). This might include ancestry; it might include health-related information.
   2. Non-monetary incentives (OR=1.72)
   3. shorter e-questionnaires (OR=1.73)
   4. A statement that others had responded (OR=1.52)
   5. A more interesting topic (OR=1.85)
   6. Using a white background (OR=1.31)
   7. Personalized e-questionnaires (OR=1.24)
   8. A simple header (OR=1.23)
   9. Giving a deadline (OR=1.18)
   10. Including a picture in the email (OR=3.05)
   11. Using the word "survey" in the email subject line (OR=0.81)
   12. E-mail included male signature DECREASED response rate (OR=.55)
2. Postal surveys (based on 481 eligible trials)
   1. Monetary incentives (OR=1.87)
   2. More interesting questionnaire topic (OR = 2.0)
   3. Pre-notification (OR=1.45)
   4. Follow-up contact (1.35)
   5. Unconditional incentives (1.61)
   6. Shorter questionnaires (1.64)
   7. Mentioning an obligation to respond (1.61)
   8. Non-monetary incentives (1.15)
   9. An assurance of confidentiality (1.33)
   10. University sponsorship (1.32)
   11. Questions of a sensitive nature DECREASED response rate (OR=.94)

Return of Results

Pathogenic variants for Breast Canncer and Alzheimer's on the exome chip array. Click Here