EAH Meiosis Apps

Background: In October, 2013, I created a mirror site for Stahl Lab Online Tools. In Spring and Summer, 2020, I was asked to modify the site and also, separately, asked to help with several analyses including one using the Papazian method. I found errors in the coding for the Papazian method. One error is that the method does not return meaningful results if the observed tetratype frequency is greater than 2/3. However, the code did return a result without warning. Additionally, there was a typo in the code for the approximate sampling variance of the NPD ratio. There is a term with \(4E^2\) in the denominator. It was coded as "/4*E^2" without parentheses so it divided by four but mutiplied by \(E^2\). This is just a typo, but it was not caught until July 15, 2020. Additionally, researchers needed output that was not provided in the tools.

In Spring 2020, I learned about Shiny Apps. Shiny had been around for at least five years, but I had been occupied chairing the Department of Mathematics at Indiana University and was in the process of transitioning to chairing the Department of Statistics. Shiny is a way of providing rather nice online tools that use R to return analyses of various types. This is much nicer and easier than writing PHP code to conduct similar analyses. The apps can use the full power of R, all of its packages, and its graphical capabilities to return results if one just learns a little bit about reactive programming (the user changes something or hits "calculate" and the output changes reactively.)

Consequently, I rented a virtual server from GoDaddy, installed security certificates, R, a shiny-server, apache2, learned how to configure all of that, and started writing shiny apps. Links to the apps with a brief description are given as a list below. On the page with the app, there is a detailed description along some references and a link to the code for the app itself.

Disclaimer: The code may contain bugs. I will try to test the code against independent code or known results and eliminate major bugs before publishing apps here. The code will be made available for you to read. If you catch additional bugs, please let me know at mail@elizabethhousworth.com.

  • The Better Way -- Goodness of Fit test for whether the data is compatible with a model of no interference.
  • Perkins Formula -- Map distances and standard errors using the Perkins formula, including testing under two conditions.