Software Brings Machine Learning to Microscopy

January 26, 2011 – 2:44 am

The sight of a researcher sitting at a microscope for hours, painstakingly searching for the right cells, may soon be a thing of the past, thanks to new software developed by scientists at the European Molecular Biology Laboratory (EMBL) in Heidelberg, Germany. Published in the science journal Nature Methods (“Micropilot: automation of fluorescence microscopy–based imaging for systems biology”), the new software rapidly learns what the researcher is looking for and automatically performs complex microscopy experiments.

Called Micropilot, the software analyses low-resolution images taken by a microscope and, once it has identified a cell or structure the scientists are interested in, it automatically instructs the microscope to start the experiment. This can be as simple as recording high-resolution time-lapse videos or as complex as using lasers to interfere with fluorescently tagged proteins and recording the results.

The developers claim that the software can generate more data, faster. In just four nights of unattended microscope operation, it detected 232 cells in two particular stages of cell division and performed a complex imaging experiment on them, whereas an experienced microscopist would have to work full-time for at least a month just to find those cells among the many thousands in the sample. They say that with such high throughput, Micropilot could easily and quickly generate enough data to obtain statistically reliable results, allowing scientists to probe the role of hundreds of different proteins in a particular biological process.

The Micropilot software is available as open source code.

Source: European Molecular Biology Laboratory

Related Posts Plugin for WordPress, Blogger... Yvonne Klöpping

Tags: , , , ,

Bookmark and Share
  1. 1 Trackback(s)

  2. Jan 27, 2011: Tweets that mention medtechinsider » Blog Archive » Software Brings Machine Learning to Microscopy --

Post a Comment