Let's start today's post with a question: Do you know what melissopalynology is? Perhaps not everyone can answer this question with a clear yes, of course. In short melissopalynology is the study of pollen contained in honey in particular to determine the source of the pollen. Not only is this used to determine the food sources and habitat range of bee colonies but also to combat fraud and mislabeling of honey. There are large value differences between honey derived from particular plants. Some monofloral honey can be much more valuable than honey derived from a mixture of plant species. Also, in the light of current threats to bee health it is crucial to determine if there is sufficient foraging habitat available as it is known that pollen diversity and quantity affect disease tolerance and longevity in bees.
Traditionally honey bee collected pollen is identified through morphological features revealed with light microscopy. However, this requires considerable skill and experience and in some plant families (Campanulaceae, Lamiaceae, Poaceae, Rosaceae) this is hampered by either a lack of characteristic features or high levels of morphological variation. It comes to no surprise that colleagues are looking into DNA-based alternatives such as metabarcoding. DNA barcoding of pollen collected by bees has been successfully done before but that pollen was directly sampled from the animals.
Researchers from the UK compared DNA metabarcoding utilizing rbcL and 454 pyrosequencing with microscopic analysis for nine honey samples provided by domestic beekeepers in Wales and England (UK). Wales was the first nation in the world to have barcoded all of its native flowering plants and conifers, enabling the identification of any plant species in Wales from the tiniest fragment of leaf, seed or pollen grain. In total 98% of the Welsh native flora (1143 species) has DNA barcodes for rbcL and 90% for rbcL and matK. Such a reference library provides perfect conditions for a metabarcoding approach. I am happy to report that Wales is not alone anymore as DNA barcoding of Canada's flowering plants is nearly completed as well.
The study results: There was 92% correspondence for the plant taxa that had an abundance of over 20%. However, the level of similarity when all taxa were compared was lower, ranging from 22–45%, and there was little correspondence between the relative abundance of taxa found using the two techniques. DNA metabarcoding provided much greater repeatability, with a 64% taxa match compared to 28% with melissopalynology. DNA metabarcoding has the advantage over melissopalynology in that it does not require a high level of taxonomic expertise, a greater sample size can be screened and it provides greater resolution for some plant families. However, it does not provide a quantitative approach and pollen present in low levels are less likely to be detected.
It is true that at this point metabarcoding is not suitable if one wants to determine pollen quantity and it a known fact that it fails to detect the rare species of a sample. There is some debate around those issues with some people claiming that especially PCR-based metabarcoding assessments can only provide presence-absence metrics while others think that this depends on the HTS platform used.
Anyway, I couldn't agree more with their final conclusion - DNA metabarcoding can be easily and widely used to investigate floral visitation in honey bees and can be adapted for use with other insects. It provides a starting point for investigating how we can better provide for the insects that we rely upon for pollination. I might add that the study clearly shows that we are capable to determine food sources of honey bees through analyzing pollen and other plant remnants in honey for food authentication and safety especially in countries where a large portion of the floral biodiversity has been barcoded.