Pairwise comparisons in linear models and p-values of contrasts

MetaboAnalyst
,
1

Hi,

I am running an analysis among 3 treatments (d5 vs. d10 vs. d15), however I am not being able to visualize the p-value of specific contrasts comparisons (for example, results from d5 vs. d10, or results from d5 vs. d10).

I understand that 238 features are statistically different (as shown in picture 1), but when I try single analysis comparing two factors, I am not able to detect any differences.

1
11544×875 49.2 KB

Then, when I look the data, I am not able to separate the results from the first contrast comparison (d5 vs. d10) and the second contrast (d5 vs. d10). I only have 1 general p-value (as shown in picture 2).

2
2947×726 26.7 KB

Question 1: Is it possible to obtain the p-values for each contrast?
Question 2: Is it possible to obtain results from all comparisons? For example, d5 vs. d10, and d5 vs. d15, and D10 VS. D15 (this one is not included because of the reference group item).

Thank you.

2
  • The result you obtained is the ANOVA style as you specified;
  • You can directly compare any two groups of interest by using Reference and Contrast;
  • Performing pair-wise (and post-hoc) analyses are possible, which is similar to batch analysis and is more suitable with the R package