Verily, the Alphabet Inc.'s research firm previously known as ‘Google Life Sciences’ measures protein profiles to discover new disease biomarkers in a series of its programs. The programs include ‘The Project Baseline Health Study’, which has gained the support of Pfizer and other drug manufacturers.
For achieving accurate results, the scientists at Verily are using specific form of protein mass spectrometry for protein identification and estimation. The major problem in this research is the approach used, which is called data-independent acquisition. This approach depends on scientific experiments to determine the spectral libraries. Moreover, the process of developing these libraries was both resource and time-intensive.
The data science experts at Verily have now collaborated with the scientists from Google and Max Planck Institute of Biochemistry to enhance the process of exploring the usage of machine learning tools. The research is supported by a belief that computations can accelerate the development of spectral libraries and thereby improve the study of protein datasets.
Substantial developments in the molecular biology sector and advances in various laboratory procedures are likely to support the growth of the global diagnostic biomarker market in the coming years. Additionally, rise in implementation rate of the biomarkers in diagnosis has increased the use of biomarkers in the diagnosis of various diseases such as cardiac, genetic, cancer and other immunological disorders. Moreover, for medical investigations, the biomarkers can be a powerful and dynamic tool to investigate the wide range of neurological disorders.
Recently, Verily has developed a DeepMass tool that is accessible as a service using the Google Cloud Machine Learning Engine. By launching this tool, Verily intends to help the inventors of diagnostics to characterize disorder that is relevant to the protein profiles. In addition, the tool has expanded Verily’s life sciences research operation that includes blood glucose tracking deal with Dexcom, a robotic surgery partnership with Johnson & Johnson, and host of other enterprises.
Verily has explained the program and its outcomes in a paper published in the Nature Methods, which is a science methodology journal publishing laboratory techniques and method papers in the life sciences. According to the authors of this paper, DeepMass tool has offered "markedly improved" results than MS2PIP, a prevailing estimation model used to predict peptide mass spectra. Moreover, the tool has attained a score that is just less than the hypothetical maximum imposed by the variability of the dataset.
According to Verily, the more accurate peptide mass spectra estimates are, greater are the chances of determining hard-to-find connections between disease and biomarkers. Additionally, according to the clinical data, the new tool is capable of covering almost double of the other known biomarkers.