Appropriate regression analysis for relationship of WGCNA modules with clinical traits

Question

[seta]

Dear all,

I wanted to do a linear regression analysis to find the relationship between some modules obtained by WGCNA and some clinical traits, such as diagnosis age, survival, etc. However, I found that there is no linear relationship between the dependent variable (eigengen) and age and also survival time. As I searched, quantile regression could be a suitable choice in this situation. However, I am not sure about it because I could not find enough papers in the field that used quantile regression. Can you share your thoughts?

Here is the scatter plot of one of the modules (eigengen) with age, considering this, I am thinking of using polynomial regression. Can you please help me to select the appropriate method for this analysis?

Thanks

[edoardofilippi1998]

I am also kinf of a beginner in this topic, and i am actually facing a simialr problem at the moment, so I am not sure I can help you, but I will try.

I have a couole of questions,

what is the goal of your analysis? Why are you trying to fit a model to the modules trait? If you want to know the significance of each module for the corresponding trait to my understanding you compute a correlation between eigengene and trait. And this is defined as the gene significance.

What is the eigengene in your plot? I mean a eigengene does not have one value but it is a vector in the gene expression space.

[seta]

my main goal is to find possible differences among 4 groups of cancer. I found non-preserved modules among these 4 groups and wanted to recognize their potential association with some clinical traits adjusted for other covariates via regression analysis within each group. Eigengene in my plot is what you defined. it's the first principal component of the module of interest, which is one value for each sample.