Skip to main content

Advertisement

Fig. 5 | Big Data Analytics

Fig. 5

From: Identification of disease-distinct complex biomarker patterns by means of unsupervised machine-learning using an interactive R toolbox (Umatrix)

Fig. 5

U*-matrix representations of lipid marker plasma concentration patterns (d = 25 markers, Fig. 1) observed in n = 200 blood samples. The figure has been obtained using a projection of the data points onto a toroid grid of 50 × 80 = 4000 neurons where opposite edges are connected. The U*-Matrix was colored as a geographical map with brown (up to snow-covered) heights and green valleys with blue lakes. Valleys indicate clusters and watersheds indicate borderlines between different clusters. The dots indicate the so-called “best matching units” (BMUs) of the self-organizing map (SOM), which are those neurons whose weight vector is most similar to the input. A single neuron can be the BMU for more than one data pint or subject, hence, the number of BMUs may not be equal to the number of subjects as in the present case. a: top view on the U matrix showing two distinct regions (clusters) on the left and right of the white “mountain range” in the middle. The BMUs are colored neutrally (grey). b: The BMUs were differently colored to analyze the distribution of subjects across the cluster structure of the data space. When the group membership to either the Parkinson patients (green dots) or the healthy subjects (blue dots) is projected onto the U*-matrix, it becomes clear that the separate clusters perfectly coincide with the diagnostic classification of the subjects. c: The cluster structure was destroyed by permutation of the data, resulting on the U-matrix display in a clearly absent cluster structure with data from Parkinson patients mixed with data from healthy subjects and no clear “mountain ranges”. The figure has been created using the R software package (version 3.4.0 for Linux; http://CRAN.R-project.org/ [24]). Specifically, the figures displaying geographical map analogies have been created using our R package “Umatrix” (https://cran.r-project.org/package=Umatrix)

Back to article page