Epsilon Archive for Student Projects

Abstract

High content screening (HCS) has potential to transform many biological fields, ranging from drug discovery to gene function discovery. HCS with time lapsed microscopy provide valuable insight information about live cells experiments that are usually lost during manual end point experiments. By means of novel bioinformatics algorithms, huge amount of phenotypic data might become available by these techniques which can be used to understand effects of chemical compounds on the cells and profile phenotypically both cell line and chemical compounds. The resultant data can also be compared with other experiments to find out the efficiency and affectivity of the different compounds under same conditions. Recent results also demonstrate that phenotypic profiles can be used to infer specific gene perturbations.
In this thesis, novel algorithms for such phenotypic profiling were implemented and demonstrated to be very useful revealing unknown kinetic information regarding two proteasome inhibitors (Bortezomib and CB3) as well as about cell clump formation during cell line growth on honeycomb nanoculture plates. The novel algorithms include specialized solutions both for phase contrast microscopy and fluorescent microscopy and are based on the publicly available cell image processing package Cell Profiler from Broad Institute.