Search

We propose a change-point detection method for large scale multiple testing problems with data having clustered signals. Unlike the classic change-point setup, the signals can vary in size within a cluster. The clustering structure on the signals enables us to effectively delineate the boundaries between signal and non-signal segments. New test statistics are proposed for observations from one and/or multiple realizations. Their asymptotic distributions are derived. We also study the associated variance estimation problem. We allow the variances to be heteroscedastic in the multiple realization case, which substantially expands the applicability of the proposed method. Simulation studies demonstrate that the proposed approach has a favorable performance. Our procedure is applied to an array based Comparative Genomic Hybridization (aCGH) dataset.

Brain tumors are the leading cause of cancer-related deaths in children. Tailored therapies need preclinical brain tumor models representing a wide range of molecular subtypes. Here, we adapted a previously established brain tissue-model to fresh patient tumor cells with the goal of establishing3D in vitro culture conditions for each tumor type.Wereported our findings from 11 pediatric tumor cases, consisting of three medulloblastoma (MB) patients, three ependymoma (EPN) patients, one glioblastoma (GBM) patient, and four juvenile pilocytic astrocytoma (Ast) patients. Chemically defined media consisting of a mixture of pro-neural and pro-endothelial cell culture medium was found to support better growth than serum-containing medium for all the tumor cases we tested. 3D scaffold alone was found to support cell heterogeneity and tumor type-dependent spheroid-forming ability; both properties were lost in 2D or gel-only control cultures. Limited in vitro models showed that the number of differentially expressed genes between in vitro vs. primary tissues, are 104 (0.6%) of medulloblastoma, 3,392 (20.2%) of ependymoma, and 576 (3.4%) of astrocytoma, out of total 16,795 protein-coding genes and lincRNAs. Two models derived from a same medulloblastoma patient clustered together with the patient-matched primary tumor tissue; both models were 3D scaffold-only in Neurobasal and EGM 1:1 (v/v) mixture and differed by a 1-mo gap in culture (i.e., 6wk versus 10wk). The genes underlying the in vitrovs. in vivo tissue differences may provide mechanistic insights into the tumor microenvironment. This study is the first step towards establishing a pipeline from patient cells to models to personalized drug testing for brain cancer.