A transcriptome-based classifier to determine molecular subtypes in medulloblastoma.

摘要: Medulloblastoma is a highly heterogeneous pediatric brain tumor with five molecular subtypes, Sonic Hedgehog TP53-mutant, TP53-wildtype, WNT, Group 3, and 4, defined by the World Health Organization. The current mechanism for classification into these subtypes through use of immunostaining, methylation, and/or genetics. We surveyed literature identified number RNA-Seq microarray datasets in order to develop, train, test, validate robust classifier identify medulloblastoma transcriptomic profiling data. have developed GPL-3 licensed R package Shiny Application enable users quickly robustly classify samples using utilizes large composite dataset (15 individual datasets), an study, dataset, gene ratios instead expression measures as features model. Discriminating were limma classified unweighted mean normalized scores. utilized two training applied 15 separate datasets. observed minimum accuracy 85.71% smallest maximum 100% four overall median 97.8% across datasets, majority misclassification occurring between 3 4 subtypes. anticipate this subtype will be broadly applicable cancer research clinical communities.