+ Reply to Thread
Results 1 to 2 of 2

Thread: Which Test: selecting two predictors of disease outcome

  1. #1
    Points: 11, Level: 1
    Level completed: 21%, Points required for next Level: 39

    Thanked 0 Times in 0 Posts

    Which Test: selecting two predictors of disease outcome


    I'm wondering which test I should use to analyzed my data.
    What I have are two groups: progressing and non-progressing patients; and I want to see if this can be predicted by the expression values of any two genes, I want to analyze every possible pair of the genes I have tested, to see which pair is the most significant.
    Any thoughts on which test I should use, also how to run it?

    Thanks for your help everyone

  2. #2
    Points: 501, Level: 10
    Level completed: 2%, Points required for next Level: 49

    Thanked 3 Times in 3 Posts

    Re: Which Test: selecting two predictors of disease outcome

    First: select genes that are of potential interest. If you have a very large number of genes, use a threshold for the CV (coefficient of variance = variance filtering) or a minimal mean fold change between the groups to pre select genes. Next, test the individual genes for difference between the groups (t-test or mann whitney u test depending on the sample size).

    Next: combine genes with relevant and significant differences between the groups in a logistic regression model in pairs. Dependent: phenotyp (0 non progression, 1 progression). Regression coefficients will be the OR for progression, i.e. a quantification of the predictive strength of one gene for the progressive phenotype corrected for the influence of the second gene. <1 = predictive for non progression >1=contributes to progressive phenotype.

    Best, Martin

+ Reply to Thread


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts

Advertise on Talk Stats