How is the the heatmap clustering figure normalized?

MicrobiomeAnalyst
,
1

Picture1
Picture1671×621 321 KB

Hello, for the heatmaps generated using microbiome analyst (MDP) > Clustering & correlation > heatmap clustering, it gives me the figure as this one above. In the normalization step when preparing data I rarefied my data to the minimum library size.

My question is related to the figure obtained - What is the meaning of the colors in the legend (like what -4 to +4 means)? From these numbers in the color legend I understand some other kind of normalization besides rarefaction was performed to generate the figure, what is the normalization used for figure?

Many thanks
Thaisa

2

Hello,

Thanks for posing the question.

The color of the legend represent the value scaled for each feature (after your rarefying) among the samples.

We will add the options to make the background normalization transparent soon.

Hope this helps.

Yao

3

Many thanks for replying so quickly. My question though is what is the method used for scaling features?
I commonly see features scaled based on Z-score in heatmaps, is that the case? Z-scores are computed on a feature-by-feature (row-by-row) basis by subtracting the mean and then dividing by the standard deviation. Can you confirm if that is what was done?

Thaisa

4

Hello, we currently provide both the feature scale and sample scale. The method is as you described, namely subtracting the mean and then dividing by the standard deviation.