Figure 2.

Random forest classification of SPTB placenta samples that underwent metabolomic analysis listed by variable importance, and colored-coded by biochemical class (n = 12 preterm and 12 control placenta samples). Random forest classification is an unsupervised analytic tool for ranking the importance of variables within a data set. The X-axis represents the mean decrease accuracy (MDA), a calculated metric of biochemical importance in characterizing the overall differences between experimental groups. A higher MDA indicates greater importance in characterizing the differences between pre-term and control placentas.