More than mesolectic: Characterizing the nutritional niche of Osmia cornifrons
Data files
Oct 13, 2023 version files 26.31 KB
-
File_1_tabs.txt
7 KB
-
File_2_tabs.txt
7.58 KB
-
File_3_tabs.txt
9.24 KB
-
README.md
2.49 KB
Abstract
Characterizing the nutritional needs of wild bee species is an essential step to better understanding bee biology and providing suitable supplemental forage for at-risk species. Here, we aim to characterize the nutritional needs of a model solitary bee species, Osmia cornifrons (Radoszkowski), by using dietary protein-to-lipid ratio (P:L ratio) as a proxy for nutritional niche and niche breadth. We first identified the mean target P:L ratio (~3.02:1) and P:L collection range (0.75 to 6.26:1) from pollen provisions collected across a variety of sites and time points. We then investigated the P:L tolerance range of larvae by rearing bees in vitro on a variety of diets. Multifloral and single-source pollen diets with P:L ratios within the range of surveyed provisions did not always support larval development, indicating that other dietary components such as plant secondary compounds and micronutrients must also be considered in bee nutritional experiments. Finally, we used pollen metabarcoding to identify pollen from whole larval provisions to understand how much pollen bees used from plants outside of their host plant families to meet their nutritional needs, as well as pollen from individual forager bouts, to observe if bees maintained strict floral constancy or visited multiple plant genera per foraging bout. Whole larval provision surveys revealed a surprising range of host plant pollen use, ranging from ~5–70% of host plant pollen per provision. Samples from individual foraging trips contained pollen from multiple genera, suggesting that bees are using some form of foraging decision-making. Overall, these results suggest that O. cornifrons have a wide nutritional niche breadth, but while pollen P:L ratio tolerance is broad, a tolerable P:L ratio alone is not enough to create a quality diet for O. cornifrons, and the plant species that make up these diets must also be carefully considered.
https://doi.org/10.5061/dryad.5x69p8d9c
Description of the data and file structure
File 1 - Protein and lipid concentrations for O. cornifrons pollen provisions across sites and time periods in Centre Country, Pennsylvania. This data was used to create Figures 1 and 2. Rows are sample names and columns are protein concentration (ug/mg), lipid concentration (ug/mg), protein:lipid ratios, and standard deviation values for the triplicate average when calculating concentrations from machine readings.
File 2 - Plant genera present in O. cornifrons pollen provisions, determined via pollen meta-barcoding. Rows are samples and columns are plant genera. Values are proportions of the total number of sequencing reads. This data was used to create Figures 3 and 4.
File 3 - Larval survival, adult survival, number of days to cocoon completion, and number of days to adulthood/pupation of bees reared on altered multifloral pollen diets (tab 1), unaltered multifloral pollen diets (tab 2), and unaltered single-source (SS) pollen diets (tab 2). This data was used to create Figures 5, 6, 7, S1, and S2. Diet is the dietary treatment group, Plate is the well plate ID bees were on, Position is the plate well bees were in, DOY is the day of year an individual perished (only for altered multifloral pollen diets), Adult status is the adult mortality status of the individual, and Larval status is the larval mortality status of the individual. A value of 1 indicates that an individual perished before reaching that stage and 0 indicates survival. DOC is the number of days an individual took to complete a cocoon (only tab 2) and DOP is the number of days an individual took to reach adulthood (only tab 2).
File 4 - Code used to run data analyses in R.
Sequencing data access
Sequences used to make data file 2 can be accessed via NCBI Sequence Read Archive under Bioproject SUB13689831.
Code/Software
Custom Python scripts were used to process meta-barcoding sequences and can be found at https://github.com/RTRichar/SimpleSequenceClassification. Code used to curate the databases used can be found at https://github.com/sbresnahan/metacleaner.
Figures 1, 2, 3, 4, S1, and S2 were created in R (version 4.1.3). Code used is available in data file 5. Figures 5, 6, and 7 were created in JMP (version 16.1) and do not have available code.