Skip to main content
Fig. 2 | Microbiome

Fig. 2

From: Proportion-based normalizations outperform compositional data transformations in machine learning applications

Fig. 2

The low abundance/high variance filtered trees are less cluttered than the unfiltered originals in the Jones dataset. The unfiltered SILVA LTP reference tree (A) shows more taxa than the filtered version (B). The UPGMA (C) and IQTREE (E) algorithms are both greedy and thus the trees are too dense and were only used after low abundance/high variance filtering (D and F, respectively)

Back to article page

Microbiome

ISSN: 2049-2618

Contact us