Currently, Biogenity is developing a product for a big data-validation tool for experimental biomarkers by generating a better overview and understanding of diseases and animal as well as cellular models. Machine learning and deep learning are used to bring unique patterns to attention from the public and in-house expression datasets. This gives a unique opportunity to discover previously unrecognized disease patterns.
We are forming a database containing an overview of all tissue wise gene and protein expressions found associated with disease and their animal as well as cellular models. We will make big data biomarker-validation more accessible for pharmaceutical companies, universities, and NGOs. Thus, providing the possibility to test the hypothesis before initiating animal, cellular or clinical experiments.
Share our vision of better biomarker-validations using existing data. If you consider donating relevant expression data, please contact us at firstname.lastname@example.org.
Big data analysis is a comprehensive process. Curating this amount of data is time-consuming if done manually, e.g. Planey and Butte reported in an article from 2013 that the curation time was 400 hours (>2 months) in an example dealing with 30 experiments. With our approach, we can deliver the data overview in approximated 7 workdays, even when including more experiments.
Our method and underlying algorithms minimize the manual labor of the curation as only the data selection process is partially supervised machine learning, otherwise, the script is unsupervised. Since datasets differ in platform, depth, and origin of samples, our algorithms will incorporate these factors and thus the normalization will account for these aspects.