Fig. 2. A heatmap of the gut microbiota of the preterm children receiving placebo.

The counts of the bacterial groups have been converted into relative abundances, then CLR-transformed for the samples, and Z-transformed for the features (taxa) so that the heatmap is distributed around 0. The y-axis shows every phylum (a), the 10 most common families in these samples (b), the 10 most common genera in these samples (c), and the 20 most common ASVs in these samples (d) associated with the most specific accuracy achievable by using the Greengenes database. The accuracy which was possible to achieve is described by letters G for genus, F for family and O for order.