For example, I hypothesis a phenotype is enhanced by the synergistic loss of function of two genes. Where the classes are 1= wild type and 2-5 are increasing subjective measures of severity of the phenotype. I wish to gauge the 'significance' of any changes to the frequency the phenotype occurrence among the genotype classes.
Ryan
Hi. I have read your question. I can help you to formulate a meaningful significance test for your scenario. Please feel free to write to me at "[email protected]" with more data (I presume you would have some tabulated data with classes, frequencies, and genotypes, right?) so I can formulate the test.
By the way, in the meanwhile download and install R into your PC.
Mohsen
To start with, you can use chi-square test or fisher exact tests to determine if there's any significant difference between your grade groups. Alternatively, consider use a logistic regression (make each grade as a predicting variable and the disease status as dependent variable). Most statistical packages (R, stata, SPSS) should be these tests for you.
You might also try something simplistic.. like a proportion test, in R, prop.test(). Say the number of scores above a critical threshold, score 2 or 3 for example. Or Fisher's Exact with the same data. But as Hao's suggested (Chi Square) is the best place to start and there are tons of free online tools as well as the stat packages recommended.
Ryan, consider the use of GLM mixed models and alelle substitution models. If normality is OK You can go ahead without problems, but if not more complex models are necessary. You can model this in R, SAS, Stata, SPSS and others.
- Badges
- Science topic
- Similar topics
- Biological Science
- Genetics