Initial refinement of data from video-based single-cell tracking

Abstract

This work proposes initial refinements of data from tracking single cells in video from many experiments and at a scale assumed large enough to be meaningful for big data analyses. The present examples of data processing are for illustration only, and caution must be exercised not to consider this contribution as a stand-alone laboratory study. The main intention is to illustrate prototypes of data refinement to help initiation of fault-tolerant big data analyses to search for causal relations in possible future large amounts of data of varying quality. The authors conjecture that computer assisted comparison between cellular behavior in many and diverse experiments can constitute a new source of information. Diversity of data sources is an argument to utilize methods employing simplicity, transparency, low-cost and sensor independence as well as independence of software or alternative online services. Indicating the potential value of simple cell positional observations (tracks), can pave the way towards contributions to already established biological databases. This will presumably help to accumulate experience from a variety of data sources and facilitate big data analyses to search for new phenotypic signatures. Data can be spin-off from special analyses. Simple perturbation of the raw data can help to check for robustness of parameters derived from it. The present example data are from tracking clonal (A549) cells during several cycles while they grow in two-dimensional (2D) monolayers. The results from its processing reflect heterogeneity among the cells as well as inheritance in their response to treatments. The present illustrations include parametrization of population growth curves, aimed to simplify computerized search for similarities in large sets of single-cell tracking data. Other types of statistics, which can promote synergy between experiments, are from temporal development of cell speed in family trees with and without cell death, correlations between sister cells, development of single cell average displacements and the tendency of clustering. These are examples of parameters for further development to utilize information in large collections of data from cell behavior.