Hay fever is an increasing problem and it is estimated that 1 in 5 people suffer from its effects. The exact species composition of pollen in the air is often unknown. I therefore explore new and more efficient ways of identifying pollen. I use automatic image recognition to relieve the time-consuming task of pollen specialists. DNA metabarcoding further aids in species recognition of plant groups that have highly similar pollen.
Pollen, hayfever, machine learning, automatic image identification, DNA metabarcoding, metagenomics
Automatic image detection and DNA metabarcoding of pollen samples
Our changing climate has severe effects on plant growth. A positive correlation has been shown between atmospheric CO2 and pollen production. Not only do many plant species show a longer flowering period, they also produce larger amounts of pollen. Apart from this, the higher temperatures allow newly introduced species to establish themselves in previously unsuitable habitats. Some of these ‘exotics’ are hayfever plants. Quantifying the contribution of pollen from many plant species to the environment is currently very hard as for some their pollen look identical to those of other species. That is why I am developing innovative tools to investigate this.
Collaborating with IBL, LIACS, University of Utrecht and University of Illinois