This week, I jumped on the opportunity to apply my skills to a different field of research.
On Tuesday, researchers working in the Stem Cells, Ageing and Molecular Physiology Unit at the
They’re using a technique involving DNA methylation: placing a chemical tag on certain genes of interest that are changed with certain environmental stimuli and measuring the resulting change in the methylation concentration. For their work, in order to calculate the percentage concentration of methylation in a sample of DNA, certain measurements of properties of experimental samples need to be calibrated against the same measurements obtained for samples of known methylation concentration. I won’t go into further detail, at risk of inadvertently exposing their methods, results, etc. – and the level of my ignorance of the subject!
After an email exchange, I realized that a lot of the data analysis they are attempting could already be swiftly accomplished with some of the python code I’ve already written. Most of my daily work involves data analysis, usually using a mixture of Python and C++ within the Ubuntu or Fedora OSs. The snippets of code from programs I’ve developed to solve astrophysical problems can, in principle, be applied to the analysis of any dataset. After all, maths, statistics and scientific computing are essential tools for any area of research. Their task of finding the calibration between the observed properties and the methylation calibration is very similar to some code I wrote for the data reduction of optical spectroscopy. In spectroscopic data reduction, the observed wavelength range of galaxy light that is incident on the CCD from the spectrograph is calibrated against identical observations taken of arc lamps, which have strong emission lines that occur at known wavelengths. Very similar stuff.
It only took a couple of evenings to adapt my existing python code to perform the task. The result: a program called MethylCal (not a very imaginative name: simply short for Methylation Calibration). The program reads in the data from an Excel spreadsheet, determines the calibration solution using the measurements of the DNA samples of known methylation concentration and then applies this calibration to the experimental DNA samples. I encountered a number of problems, which mainly arose from this being my first experience at turning a python program in Linux into a self-contained executable for Windows. I’ll shortly blog about some of these issues and solutions. Yet, the program was able to successfully perform the analysis and generate results. The program is designed to be flexible enough for the researchers at RISES-LJMU to analyse their data for all the experiments concerning different genes.
Overall, I learnt a lot from the project. Creating something that may/will be a useful tool in another field is immensely satisfying. Hence, I hope to have more opportunities to work in such interdisciplinary collaborations in the future.
Now… back to writing the NGC 891 paper!