Abstract

The recent development of high-throughput drug profiling (high content screening or HCS) provides a large amount of quantitative multidimensional data. Despite its potentials, it poses several challenges for academia and industry analysts alike. This is especially true for ranking the effectiveness of several drugs from many thousands of images directly. This paper introduces, for the first time, a new framework for producing graph theoretic descriptors and automatically ordering the performance of drugs, called fractional adjusted bi-partitional score (FABS). This general strategy takes advantage of graph-based formulations and solutions and avoids many shortfalls of traditionally used methods in practice. We experimented with FABS framework by
implementing it with a specific algorithm, a variant of normalized cut - normalized cut prime (FABS-NC), producing a semi-supervised measure. This algorithm is known to run in polynomial time and therefore can scale well in high-throughput applications. We assessed the accuracy of the resulting measure by a comparative study, which includes different implementations of the FABS framework with support vector machine(FABS-SVM) and normalized cut spectral
technique(FABS-Spectral) - in addition to other three baselines: center ranking(Center), PCA ranking (PCA) and Z-Factor. The conclusion is encouraging: FABS-NC consistently outperforms other methods and is the most stable implementation out of the three. In some cases, producing over half correctly predicted ranking experiment trials than the next best algorithm.


Keywords – High Content Screening, Graph Bi-Partition, Graph Cut, Drug Ranking.
Topics – Imaging.