Bisque: Advances in Bioimage Databases

Biological image databases have quickly replaced the personal media collections of individual scientists. Such databases permit objective comparisons, benchmarking, and data-driven science. As these collections have grown using advanced (and automated) imaging tools and microscopes, scientists need high-throughput large-scale statistical analysis of the data. Traditional databases and standalone analysis tools are not suited for image-based scientific endeavors due to subjectivity, non-uniformity and uncertainty of the primary data and their analyses. This paper describes our image-database platform Bisque, which combines flexible data structuring, uncertain data management and high-throughput analysis. In particular, we examine: (i) Management of scientific images and metadata for experimental science where the data model may change from experiment to experiment; (ii) Providing easy provisioning for high-throughput and large-scale image analysis using cluster/cloud resources; (iii) Strategies for managing uncertainty in measurement and analysis so that important aspects of the data are not prematurely filtered. 1 Challenges for Bioimage Researchers Current research in biology is increasingly dependent on conceptual and quantitative approaches from information sciences, ranging from theory through models to computational tools [6]. Ready availability of new microscopes and imaging techniques has produced vast amounts of multi-dimensional images and metadata. The introduction of new models, measurements, and methods has produced a wealth of data using image-based evidence [24]. Two notable examples of image-based studies are cellular Alzheimer’s studies and plant genetics. In a recent Alzheimer’s study, the ability to reliably detect nuclei in three dimensions was critical to quantitative analysis [25]. The use of nuclei detection also finds use in a wide range of applications such as the accurate determination of how an organism is perturbed by genetic mutations, treatment with drugs, or by injury. Additionally, nuclei centroid locations can be used for further analysis, such as cell membrane segmentation or to initialize and validate computational models of cells and their patterns. In the plant domain, technologies for quantifying plant development are underdeveloped relative to technologies for studying and altering genomes. As a result, information about plant gene function inherent in mutant phenotypes or natural genetic variation remains hidden. For example, plant scientists are trying to uncover gene function by studying seedling growth and development using high throughput image analysis [19]. In both cases, researchers dependent on images as experimental evidence face the daunting task of managing, analyzing and sharing images in addition to gaining and providing access to analysis methods and results [1]. In Copyright 2012 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE. Bulletin of the IEEE Computer Society Technical Committee on Data Engineering