I am running a signed WGCNA analysis using FPKM values from 100 maize RNA-seq samples. The samples include root, shoot, and seed tissues.

I compared two analyses. When all tissues were included, the maximum scale-free topology fit was R2 = 0.783 at power 30, and no tested power reached R2 = 0.80.

When I excluded the seed samples and analyzed root and shoot tissues only, power 10 reached R2 = 0.820 and power 11 reached R2 = 0.872.

My biological question concerns nitrate uptake in maize seedlings. Since seed tissues are biologically different from root and shoot tissues, should I run WGCNA using all tissues, or should I exclude the seed samples and analyze root and shoot tissues only?

If seeds should be excluded from the main analysis, should I analyze them separately?

I would appreciate advice on the appropriate approach and how to justify this decision in a manuscript.