In an increasing number of our experiments, we use computational image analysis to obtain quantitative and unbiased information on large numbers of individual cells. Not only does this improve the quality and objectiveness of cell biological experiments, but it is a necessary component in our work to reveal causal interactions between population-determined phenotypic properties of cells, their activities, and the molecular components that establish these interactions.
Cellclassifier was developed to allow cell biologists in our lab to do supervised machine-learning and automated classification of large numbers of cells, without the help of computer experts. It provides an intuitive and easy-to-use user interface with which a user can visualize the cells as they were imaged with the microscope, and can click on them for easy annotation. The user can also evaluate results of automated classification on the original images.
iBRAiN is an acronym for image-Based RNAi and was developed as a middle ware between image analysis software, content management of siRNA libraries, the storage of large sets of microscopy images, the distribution of computation jobs on large computer clusters, and the harvesting and visualization of obtained results. It contains numerous meta-information on genes and proteins, obtained from ontology databases and STRING. It also incorporates probabilistic algorithms to determine the true loss-of-function phenotype of a gene from measurements of multiple siRNAs targeting the same gene. Some of these algorithms have been developed by us, and will be published in the near future. iBRAIN also provides a webbrowser-compatible user interface with which the progress in computational analysis can be easily monitored for each running project.