Applications of node-based resilience graph theoretic framework to clustering autism spectrum disorders phenotypes

Abstract

With the growing ubiquity of data in network form, clustering in the context of a network, represented as a graph, has become increasingly important. Clustering is a very useful data exploratory machine learning tool that allows us to make better sense of heterogeneous data by grouping data with similar attributes based on some criteria. This paper investigates the application of a novel graph theoretic clustering method, Node-Based Resilience clustering (NBR-Clust), to address the heterogeneity of Autism Spectrum Disorder (ASD) and identify meaningful subgroups. The hypothesis is that analysis of these subgroups would reveal relevant biomarkers that would provide a better understanding of ASD phenotypic heterogeneity useful for further ASD studies. We address appropriate graph constructions suited for representing the ASD phenotype data. The sample population is drawn from a very large rigorous dataset: Simons Simplex Collection (SSC). Analysis of the results performed using graph quality measures, internal cluster validation measures, and clinical analysis outcome demonstrate the potential usefulness of resilience measure clustering for biomedical datasets. We also conduct feature extraction analysis to characterize relevant biomarkers that delineate the resulting subgroups. The optimal results obtained favored predominantly a 5-cluster configuration.Electronic supplementary materialThe online version of this article (10.1007/s41109-018-0093-0) contains supplementary material, which is available to authorized users.

Connecting phenotype to genotype: PheWAS-inspired analysis of autism spectrum disorder

“…It is interesting to note that most of the optimal results (except one) were obtained with k=3. This aligns with previous evidence presented in favor of the use of a minimal connectivity parameter k in the construction of kNN graphs (Matta et al, 2018b). We compared the clustering results for the highest ranked graph per subcategory in Table 3 across all 51 phenotype features (see Table 1) that span ASD-specific symptoms, cognitive and adaptive functions, language and communication, behavioral problems, genetic indicators, and facial biomarker.…”

“…Presenting a new classification system for heterogeneous ASD instances is beyond the scope of this paper. However, to the best of our knowledge, most prior ASD clustering work (Bruining et al, 2010 ; Matta et al, 2018b ) directly clusters the samples in an attempt to uncover more homogeneous subgroups of the sample population. In this research, we demonstrate an empirical framework for using genotype and phenotype information to cluster ASD samples in a non-traditional way.…”

Applications of Unsupervised Machine Learning in Autism Spectrum Disorder Research: a Review

“…Node-based resilience (NBR) clustering uses a graph-theoretic framework to represent and split data points into groups (Matta et al, 2018). First, graphs representing the data are created by creating a node (or vertex) for each observation.…”

“…Nodes that are a part of the attack set can either be excluded from classification or can be assigned to the nearest cluster. Various measures of node-based resilience (such as Vertex attack tolerance, integrity, and tenacity) can also be calculated (see Matta et al, 2018 for an overview of these measures)…”

“…In Matta et al, 2018Matta et al, , 2680 between the ages of 4 and 17 were examined. Features included 36 normalized phenotype features (see Matta et al, 2018 for detailed list) from the Simons Simplex Collection were used for clustering. Additional features were compared between resultant clusters to validate the differences between clusters.…”