This is a test version of Biostars. For the public version, visit https://www.biostars.org.
how to adjust a phenotype according to age

Dear all,
I am running an analysis on one phenotype and many SNPs with random forest. However, I find that phenotype correlated with age. I would like to correct phenotype with ages, then run the random forest on SNPs. At this moment, I run the linear model on phenotype with age, and used the residual as the adjusted phenotype. I am wondering there anyone know some methods could be used to correct/adjust the phenotype according with ages?
Thanks and with best regards,
Xinhui

r
• 3,037 views • link
updated by Ram Ram • written by xinhui.wang xinhui.wang

Sidebar: Why is R the only tag here? Why not add a few subject matter related tags?

• 0 views • link
written by Ram Ram

i have no idea which tag shall be added? Do you some idea?

• 0 views • link
written by xinhui.wang xinhui.wang
  • random forest
  • correlation

are a couple I can glean just from the post text. Nothing in the post suggests you're using R though.

• 0 views • link
written by Ram Ram

thanks. I changed the description!

• 0 views • link
written by xinhui.wang xinhui.wang

The post still has just one tag and the text does not mention any R package or function call, so I don't see any effective change.

• 0 views • link
written by Ram Ram

How did you determine this?

However, I find that phenotype correlated with age.

• 0 views • link
written by Kevin Blighe Kevin Blighe

correlation coefficient as 0.70 ...

• 0 views • link
written by xinhui.wang xinhui.wang

Also, is it statistically significant when you model it against your outcome in a lm() or glm() ?

• 0 views • link
written by Kevin Blighe Kevin Blighe

yes, p value < 0.00000001

• 0 views • link
written by xinhui.wang xinhui.wang

I get worried when I see p-values like that from regression. What is the standard error and Estimate? Also, the confidence intervals for the odds ratio? - A: Polygenic Risk Scores: Odds ratio or Beta-coefficient?

In any case, I don't believe there is any right or wrong answer. Also, I cannot see your data and can only ask you questions here. You may want to read this: Are your covariates under control? How normalization can re-introduce covariate effects

• 0 views • link
written by Kevin Blighe Kevin Blighe

0 answers

No answers yet.

Log in to answer this question.

More posts like this