|Math may help in a deeper understanding of stem cells|
Why do cells do what they do? Why does one cell become cancerous while its neighbor remains normal? What makes a stem cell ultimately change or differentiate into a nerve cell in the brain, rather a pancreatic beta cell? What influences those decisions?
The fact is, even two cells from the same tissue can show subtle distinctions, such as having a gene turned on in one and that same gene turned off in its sibling. Finding these differences could improve our understanding of disease and help delineate the signals that make stem cells choose one path over another.
Scientist have used a variety of methods—from microscopy to genomics—to understand what makes similar cells go in different directions. Now they appear to have a new tool: mathematics.
In a paper published in January in the journal PNAS, researchers from the Technische Universitaet Muenchen, the Helmholtz Zentrum Muenchen and the University of Virginia describe a statistical approach to illuminate the inner workings of single cells.
The method takes an existing technology, RNA sequencing, and makes it more precise. As the middleman between DNA (the blueprint) and proteins (the workers) RNA is an excellent target to help us understand what’s going on inside a cell.
But scientists have had a quandary. Analyzing single cells produces such a small sample that it can be difficult to separate good data from bad. On the other hand, using a larger sample, say ten cells, produces a better signal to noise ratio but can obscure the small but important variations between single cells.
The team discovered that the way to overcome the noise, and find the signal, was to start with a larger group of cells. The scientists then used a form of statistical analysis called maximum likelihood inference to isolate the subtle variations between single cells in the group, with excellent results.
To test their math, the researchers targeted a number of rarely expressed genes including PIK3CD, which is associated with cancer proliferation. Though most cells in the samples did not express this gene, or expressed only small amounts, the method successfully isolated groups of one or two cells with higher PIK3CD concentrations.
This new approach gives researchers the best of both worlds—the accuracy of larger samples and the unique information provided by single cells.
We are funding a training program that employs similar methods – using mathematical formula and computer programs – to help prepare a new generation of stem cell scientists.