Examples of problems that can be solved using data science:

  • Many confidential examples of predicting bad or inconsistent products based on big data. Big data consisting of e.g., quality of ingredients, cleaning regimes, processing parameters.
    Statistics, multivariate analyses, machine learning, manual curation, and follow-up experiments are typical necessities to arrive at those factors that causally affect product quality.

  • Identify from big data an essential bone mineralization factor in a human Mendelian disorder. Publication.
    Statistics, clustering and manual curation were used to integrate genomics and glycomics data to find this causally related bone mineralization factor.

  • Predict if a person has active / latent or no Tuberculosis based on big data. Publication.
    Random Forest Machine Learning, statistics and modelling applied to whole blood transcriptome to derive a predictive gene expression signature.

  • Screening big data to find toxin genes that associate to human colorectal cancer. Publication.
    Random Forest Machine Learning to pinpoint toxin genes that associate to human colorectal cancer from meta-transcriptomics data.

  • Finding metabolites that plant predators don’t like. Publication in preparation.
    Based on statistics and machine learning find metabolites that anti-correlate to presence of plant predator insects with the aim to find that metabolite that deters a particular plant predator.