Counting Pictures of Cell that Count: A Biotiff-Based Strategy for Indexing Cell Images and Associated Metadata in Large Series of Digital Microscopy-Based Biophysical Result

Abstract

Last year we described how BioTIFF code (www.biotiff.org) enables all metadata needed to interpret a given digital image region to be embedded within the same file structure used to store that image (Biophys J. 96: 30a). Here we show how that metadata can then be indexed so that Google-like queries can be performed on any set of BioTIFF files. Indexing can be done either on a single local set of BioTIFF images or a distributed set by using a distributed file system. This then creates a distributed index that has many interesting applications for exploring image relationships across different subsets of images within large sets. For example, single cell (or single molecule) responses measured using digital microscopy-based biophysical methods that relate cellular mechanism to light-based surrogate measures can then be linked to a specific cell in a specific field-of-view of a specific sample. The replication of such results over time in the same lab or in a distributed manner by multiple labs can allow for collaborative replicable science to anchor the evolution of shared understanding of cell physiology mechanisms. We will demonstrate how the approach can be implemented using a plurality of commodity gear meeting certain minimum standards and open source imaging, indexing and search software. This level of transparency and annotation of experimental detail allows for differences in experimental conditions between experiments and labs to be accommodated in open collaborative interpretation of biophysical data.