Predictive markers of honey bee colony collapse

Abstract

Across the Northern hemisphere, managed honey bee colonies, Apis mellifera, are currently affected by abrupt depopulation during winter and many factors are suspected to be involved, either alone or in combination. Parasites and pathogens are considered as principal actors, in particular the ectoparasitic mite Varroa destructor, associated viruses and the microsporidian Nosema ceranae. Here we used long term monitoring of colonies and screening for eleven disease agents and genes involved in bee immunity and physiology to identify predictive markers of honeybee colony losses during winter. The data show that DWV, Nosema ceranae, Varroa destructor and Vitellogenin can be predictive markers for winter colony losses, but their predictive power strongly depends on the season. In particular, the data support that V. destructor is a key player for losses, arguably in line with its specific impact on the health of individual bees and colonies.

Figure 1. Colony strength and V. destructor natural mite fall.

A) Colony strength (Y-axis) by sampling month in 2007/2008 (X-axis) for the colonies that died during winter (DC, black boxes) and the ones that survived (SC, grey boxes). P value is indicated (Mann-Whitney test): **P<0.01. Number of live colonies is given (N). B) Natural average V. destructor mite fall in 2007/2008 on the hive bottom boards per day per colony (Y-axis) for the two groups Dying Colonies (DC) and Surviving Colonies (SC) over the experimental period. (X-axis) Infestation levels [average ± SD] are shown for each month and timing of treatments (OA = Oxalic Acid, FA = Formic acid) is indicated by arrows (n.v. = no value).

Figure 2. Seasonal variability in Summer, Fall and Winter.

A) Pathogens B) Gene expression profiles C) V. destructor. The Y-axis shows the relative quantities except for DWV, where it shows equivalent genome copies and the X-axis displays the groups Dying Colonies DC (N = 13) and Surviving Colonies SC (N = 16). Significant differences (Two tailed t-test) are indicated with * = P<0.05 and ** = P<0.01.

Figure 3. Overall seasonal variability of colonies (N = 29) from Summer to Winter.

A) Pathogens B) Genes C) V. destructor. The Y-axis show the relative quantities except for DWV, where it shows equivalent genome copies [log10]. The seasons summer, fall and winter are shown on the X-axis. (Mann-Whitney test with Bonferoni correction,* P<0.05; **P<0.01; ***P<0.001).

Figure 4. Spearman rank correlations for the different variables who showed significant P values (P<0.05, indicated in bold) for the three seasons summer, fall and winter (N = 29 colonies).

The variables shown are: Hym = Hymenoptaecin, Vg = Vitellogenin, Eater, NCer = Nosema ceranane, DWV, BQCV and Vd = Varroa destructor.