High Content and Throughput Drug Discovery

Abstract

1.1 The marriage of ‘high throughput’ and ‘high content’ While the pharmaceutical industry innovation crisis draws much debate (Kaitin & DiMasi, 2011; Macarron et al., 2011; Munos, 2009; Paul et al., 2010), there remains little consensus on how to cohesively deliver value throughout the drug development pipeline (Fig. 1). This chapter considers some of these issues in the context of a growing field for computational biology: drug discovery high throughput screening (HTS). HTS is the approach of rapidly studying physical, chemical, biological and genetic perturbations on the scale of of tens of thousands per day. Today we are faced with ultra-HTS daily screen rates of hundreds of thousands, in part thanks to the continued development of technologies such as micro-fluidics (Agresti et al., 2010). As a discovery tool, it traces it roots back over twenty years (An & Tolliday, 2010), however it is the more recent improvements in cell culture technique with the potential for multivariate output such as gene expression that brings it into the domain of high content computational biology. With this maturation of cell-based assays we also notice an increased focus on statistical rigour, analytical integration, and the apparent user-driven plateau in miniaturisation (Mayr & Bojanic, 2009). Rather than being faced with a continued improvement in simple assay throughput, these suggest a growing role for more data-rich high content HTS (hcHTS)1. Despite the implicit gains, there exists a notable and growing antipathy towards many ‘big data’ approaches as discovery tools. Much publication has refocused on data quality versus quantity, with some doubting the impact of high throughput science altogether (Douglas et al., 2010; Macarron et al., 2011; Mayr & Bojanic, 2009). There persists the very real hurdle of experimentalists and team leaders struggling with the interpretation, integration, and decision making based on such data. As a concern routinely witnessed in post-genome era science, it is doubtful that the blame rests primarily with problem-specific methodology. In this chapter, the need for screens to bemore decision-centric and transparent across disciplines is proposed. The aim here is not just to provide the reader with specific tools that are likely to rapidy become dated, but introduce the scope and opportunities in drug screening science.