which genes are up-down regulated from Anova?

Question

[akutasame]

Hi,

Is there any way to understand which genes are up or down regulated just from Anova or from linear model coefficients?

Thanks in advance

[akutasame]

Thank you Mr. Devon,

I m really wondering if I can understand up or down regulated genes from Anova coefficients just like the following. Is that coef(lv) result giving any hint for regulated genes?

            Df Sum Sq Mean Sq F value   Pr(>F)    
A            2   1383     692   1.916    0.165    
M           15  62583    4172  11.561 1.26e-08 ***
Residuals   30  10826     361                     
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> lv
Call:
   aov(formula = pp ~ A + M)

> coef(lv)
(Intercept)     Amutant  Awild-type        M114        M115        M116        M117        M118
  24.533132   -5.888804  -13.125944    7.482282   18.960569  114.648217   -9.331482  -11.131245
       M119        M120        M121        M122        M123        M124        M125        M126
  -7.068761  -11.270026  -11.909917  -10.910291    7.137681   48.106454   -4.424028   -5.267856
       M127        M128
  82.327849   -4.887601

[Devon Ryan]

Of course, the coefficients tell you direction and magnitude of change. However, those all come with noise, which is why you also need to use the p-values.

[akutasame]

Okay, I got p values and also I used bonferroni method. Then, what should I do ?

Answer 1

[Devon Ryan]

Typically you want to know which ones are significantly up/down regulated, so you'd use an adjusted p-value in addition to the fit coefficient. In principle the fit coefficient and its confidence intervals are pretty telling, though (note that it's difficult to take multiple testing into account with this).