Korn Edward L, Habermann Jens K, Upender Madhvi B, Ried Thomas, McShane Lisa M
Biometric Research Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA.
Biotechniques. 2004 Jun;36(6):960-7. doi: 10.2144/04366BI01.
Many image analysis systems are available for processing the images produced by laser scanning of DNA microarrays. The image processing system takes pixel-level intensity data and converts it to a set of gene-level expression or copy number summaries that will be used in further analyses. Image analysis systems currently in use differ with regard to the specific algorithms they implement, ease of use, and cost. Thus, it would be desirable to have an objective means of comparing systems. Here we describe a systematic method of comparing image processing results produced by different image analysis systems using a series of replicate microarray experiments. We demonstrate the method with a comparison of cDNA microarray data generated by the UCSF Spot and the GenePix image processing systems.
有许多图像分析系统可用于处理通过对DNA微阵列进行激光扫描所产生的图像。该图像处理系统获取像素级强度数据,并将其转换为一组基因级表达或拷贝数汇总数据,这些数据将用于进一步分析。目前正在使用的图像分析系统在它们所实施的具体算法、易用性和成本方面存在差异。因此,需要有一种客观的方法来比较这些系统。在这里,我们描述了一种系统方法,通过一系列重复的微阵列实验来比较不同图像分析系统产生的图像处理结果。我们通过比较由加州大学旧金山分校的Spot图像处理系统和GenePix图像处理系统生成的cDNA微阵列数据来演示该方法。