About pmap value on heat map from permutation t-test result

Discussions
,
1

Hi everyone,

I want to ask about the pmap value on the heat map from the permutation t-test results. after I did the test, significant results (different colors) were displayed on the heat map, but after I checked the pmap value one by one, there were several columns that had significant pmap values as well, but the heat map did not read significantly (with a change in color). what is this? does anyone have knowledge or experience related to this, thank you for your help.

The left side is the result of the heatmap. The right side is after I check the pmap value one by one for each column. (yellow, significant by heatmap; green, pmap <0.05 but not read as significant on heatmap)

Screenshot 2024-07-04 172408
Screenshot 2024-07-04 1724081248×702 106 KB

Best regards,

Budi

2

Hi @budisppa777,

Could it be the case that you run the permutation t-test using either the One-tailed (-) or One-tailed (+) options?

In that case the masking of significant results will not only be the p-value, but the sign of the t-statistic.

3

Thank you for your information @Raymundo.Cassani
Now I understand, that, a significant result cannot only refer to the pmap value, but must be seen from the sign of the t-statistic as well.

4

Only for one-tailed analysis.

For two-tailed, only the p-value is taken into account to be considered statistically significant