title: Computational methods to analyze image-based siRNA knockdown screens
creator: Batra, Richa
subject: ddc-500
subject: 500 Natural sciences and mathematics
description: Neuroblastoma is the most common extra-cranial solid tumor of early childhood. Standard therapies are not effective in case of poor prognosis and chemotherapy resistance. To improve drug therapy, it is imperative to discover new targets that play a substantial role in tumorigenesis of neuroblastoma. The mitotic machinery is an attractive target for therapeutic interventions and inhibitors can be developed to target mitotic entry, spindle apparatus, spindle activation checkpoint, and mitotic exit. Thus, we performed a study to find genes that cause mitosis linked cell death upon inhibition in neuroblastoma cells.     We investigated gene expression studies of neuroblastoma tumors and selected 240 genes relevant for tumorigenesis and cell cycle. With these genes we performed image-based time-lapse screening of gene knockdowns in neuroblastoma cells. We developed a classifier to classify images into cellular phenotypes, using SVM, performing manual evaluation and automatic corrections. This classifier yielded better predictions of cellular phenotypes than the standard classification protocol. We further developed an elaborated analysis pipeline based on the phenotype kinetics from the gene knockdown screening to identify genes with vital role in mitosis to identify therapeutic targets for neuroblastoma. We developed two methods (1) to generate clusters of genes with similar phenotype profiles and (2) to track the sequence of phenotype events, particularly mitosis-linked-celldeath.    We identified six genes (DLGAP5, DSCC1, SMO, SNRPD1, SSBP1, and UBE2C) that cause mitosis-linked-cell-death upon knockdown in both of the neuroblastoma cell lines tested (SH-EP and SK-N-BE(2)-C). Gene expression analysis of neuroblastoma patients show that these genes are up-regulated in aggressive tumors and they show good prediction performance for overall survival. Four of these hits (DLGAP5, DSCC1, SSBP1, UBE2C) are directly involved in cell cycle and one (SMO) indirectly which is involved in cell cycle regulation. Functional association and gene-expression analysis of these hits indicated that monitoring cell cycle dynamics enabled finding promising drug targets for neuroblastoma cells.    In summary, we present a bioinformatics pipeline to determine cancer specific therapeutic targets by first performing a focused gene expression analysis to select genes followed by a gene knockdown screening assay of live cells.  
date: 2013
type: Dissertation
type: info:eu-repo/semantics/doctoralThesis
type: NonPeerReviewed
format: application/pdf
identifier: https://archiv.ub.uni-heidelberg.de/volltextserverhttps://archiv.ub.uni-heidelberg.de/volltextserver/14922/1/Thesis_Richa.pdf
identifier: DOI:10.11588/heidok.00014922
identifier: urn:nbn:de:bsz:16-heidok-149227
identifier:   Batra, Richa  (2013) Computational methods to analyze image-based siRNA knockdown screens.  [Dissertation]     
relation: https://archiv.ub.uni-heidelberg.de/volltextserver/14922/
rights: info:eu-repo/semantics/openAccess
rights: http://archiv.ub.uni-heidelberg.de/volltextserver/help/license_urhg.html
language: eng