Current healthcare analytics tools are limited by not being able to combine many biomarkers at any one time. In this Q&A, Henrik Kristiansen, chief executive officer at GenoKey, explains the issues and suggests some possible solutions for the future
Q. What are the problems with current healthcare analytics?
A. Current healthcare analytics miss out on some of the information hidden in the patterns of the wealth of data coming out of human studies, because they can’t look at the fine detail, and find the true impact of individual changes I amongst all the ‘noise’. For example, in a study of the genetic changes behind bipolar disorder, researchers found about 80 SNPs (single nucleotide polymorphisms) by conventional analysis, but most of these turned out be false positives.
Q. How does GenoKey’s technology help?
A. Our technology is very precise, and can find the hard-to-locate patterns in all the data. In the case of the bipolar disorder genetic data, by looking at combinations of three SNPs and testing the combinations and permutations of these using our case-control data mining workflow, we were able to find four significant clusters of SNP genotypes and their associated genotypes (clusters), and eliminate the false positives and false negatives. Our technology is also very fast, because we use the power of GPU computing.
Q. So, does your technology just work with genetic data?
A. It’s important to emphasize that our data mining technology can find relationships in many different kinds of data, not just genetic information. This includes clinical data, results from diagnostic tests, and biomarkers in blood and other body fluids.
Q. Why are blood biomarkers important?
In my opinion, blood biomarkers are very important because they can be used to predict the risk of disease before it happens, and allow doctors to work to prevent its onset, or lessen its impact, through drugs or lifestyle changes. People talk a lot about personalised medicine, but this could give us the opportunity for ‘personalised preventive medicine’.
However, looking at individual biomarkers can lead to false positives and false negatives. As an example, with colon cancer, diagnostic tests include faecal occult blood, where the stool samples are tested for blood. Not only is blood in the stool difficult to trace, but can also be harmless, and so is therefore an example of a ‘false positive’. In contrast, there may be no blood in that particular stool sample but the patient may still have cancer – this is a ‘false negative’.
As this demonstrates, using individual biomarkers, whether they are genetic or clinical, can lead to incorrect results. Combining several markers at a time, as with technologies like GenoKey’s, could lead to better stratification of patients, or higher confidence in diagnoses, for both personalised medicine and personalised preventive medicine.