An automated cell-counting algorithm for fluorescently-stained cells in migration assays

Traditionally, in vitro cell-counting methodologies consist of manual counts through use of a hemacytometer [1,2]. Generally, cell migration experiments are conducted using modified Boyden chambers, whereby the cells of interest migrate through a porous membrane and are stained for counting. Such migratory cells are commonly labelled on the membrane with a crystal violet stain [3,4], Trypan Blue dye [5], or hematoxylin [6-8], and quantified manually. Although it remains the gold standard, manual cell counting is very time-consuming and may introduce experimenter bias, thus increasing the potential for measurement errors [9].In an effort to increase efficiency and mitigate potential sources of bias/error associated with manual cell counting, a number of commercially available software suites provide automated cell counting from microscopic images. These software packages enable users to collect cell counts from random fields of view within specimens of interest. Unfortunately, these software suites contain proprietary algorithms (making them inadaptable), can be expensive, and often require high performance computers.Whole membrane quantification has been accomplished through spectrophotometric methods using absorbance microplate readers [10-12]. These microplate readers are able to detect total stained (if using dyes/stains) or fluorescence signal, where output values are a direct indication of total migrated cells. Although absorbance microplate readers allow for an expeditious analysis of migration assays, they are costly and do not provide a record of membrane images should manual confirmation or further analysis be required.In the current study, we sought to develop a feasible, valid, and reliable algorithm designed to automate cell counting using stored images from cell migration experiments. In an effort to validate our algorithm we: 1) compared automated cell counts with manual counts from a blinded experimenter; and 2) determined the effect of objective powe