Category: Automation and High-Throughput Technologies
Significant efforts have been made recently to improve data quality in high-throughput screening (HTS) technologies widely used in drug development and chemical toxicity research. However, data generated by these screening technologies are subject to several environmental and procedural spatial biases which introduce errors into the hit selection process. Although effective bias correction methods and software have been developed over the past decades, almost all of these tools have been designed to reduce the effect of additive spatial bias only. Here we describe multiplicative spatial bias. The most straightforward approach for removing multiplicative spatial bias from experimental screening data consists in the use of a logarithmic transformation of raw measurements, followed by the application of additive bias correction methods: such as B-score, SPAWN, Partial Mean Polish (aPMP), or R- score. However, a typical data preprocessing step in HTS consists of normalizing raw measurements (e.g., plate or well-wise using Z-score). This step is performed prior to the data correction and hit selection steps. Thus, the logarithmic transformation cannot be applied in many cases because the normalized measurements contain negative values. Three new statistical methods meant to reduce the effect of multiplicative spatial bias in screening technologies have been elaborated. The first method, called Non-Linear Multiplicative Bias Elimination (NLMBE), solves a system of nonlinear algebraic equations in which the unknowns correspond to spatial biases, which affect specific rows and columns of a given plate. The second method, called multiplicative PMP (mPMP), is based on a multiplicative partial mean polish procedure in which the mean of each row and each column affected by spatial bias is adjusted iteratively with respect to the mean of the unbiased plate measurements. The rows and columns affected by spatial bias can be detected by the Mann–Whitney U test. This information is required by both the NLMBE and mPMP methods. Our third method, called multiplicative B-score, is an adaptation of a 2-way median polish procedure (i.e., a multiplicative version of the traditional B-score algorithm). We have assessed the performance of the methods with synthetic and real data affected by multiplicative spatial bias, including comparisons with current bias correction methods. A general data correction protocol that integrates methods for removing both assay and plate-specific spatial biases will also be presented. We will show that our new methods for removing multiplicative spatial bias and the data correction protocol are useful for detecting and cleaning experimental data generated by screening technologies. As this protocol is of a general nature, it can be used by researchers analyzing current or next-generation high-throughput screens. The algorithms are available in the AssayCorrector R software package available at CRAN (https://cran.r-project.org/)
Iurie Caraus– PhD student, Department of Computer Science, UQAM, Montreal, QC, Canada
Department of Computer Science, UQAM
Montreal, QC, Canada
Iurie Caraus is a PhD student in the department of Computer Science at the Université du Québec à Montréal, Canada. He holds a Bachelor and a Master degree in Applied Mathematics from the Moldova State University. He is involved in the analysis of bioinformatics data related to high-throughput screening. His PhD thesis is dedicated to the development of new accurate statistical methods for detecting and minimizing systematic error in experimental HTS assays.