Article published Spectrometry
Software for medical diagnosis
A CEA LIST team working on the Computis project has developed the first software for displaying, processing and structuring mass spectrometry spectra and images. This product opens up new prospects for the pharmaceutical industry and for medical diagnosis.
Mass spectrometry is used in almost every field of science (physics, astrophysics, gas-phase chemistry, organic chemistry, assays, biology and medicine) to detect and identify molecules of interest by measuring their mass. CEA has developed a number of automatic methods to process the large volumes of data output by this type of analysis, then to classify and organise it and break it down to locate relevant information. This work is carried out as part of the European Computis project. “Unlike most software products supplied with spectrometers, the tool we have developed is not merely designed for data display and processing purposes, but also to analyse and understand this type of data, which contains a wealth of information,” explains Marie-France Robbe, a project manager at CEA LIST. Known as “fxSpectViewer”, this software uses fast, robust algorithms that take up relatively little memory space, for high-speed processing of very large volumes of data (generally from 0.5 to 5 GB per image). Until now, scientists had to study, one by one, each peak representing the molecules contained in the sample - a task that could take several days. With “fxSpectViewer”, however, it is now possible to isolate a large number of peaks providing meaningful information and characterised by their location in the image and their statistical properties. The software can also display the image representing a given peak and, for example, break it down into sub-zones. In addition, a “structuring” module has been specifically developed to structure data and break it down into relevant categories. The software has already proven its worth when it was used by CNRS and Genethon teams to reveal, in record time, a number of lipid markers of Duchenne muscular dystrophy and to verify the effectiveness of gene therapy in treating this disorder in mice. It has also been used to display how drugs bind to certain tissues. Further developments are planned. “We are currently seeking partners to develop a new version capable of accurately mapping out the lipids, peptides and metabolites found in tissues, or a version that could be used to help healthcare professionals make a diagnosis by identifying the characteristic markers of a disease, then analyse the results obtained using various imaging techniques.” This type of software opens up new horizons in many areas, such as genomics research and for the diagnosis or therapeutic monitoring of many neurodegenerative disorders and cancers.
