I am an aquatic ecologist primarily interested in the fate and importance of different sources of primary productivity in marine, lake, and estuarine foodweds. I work on these questions using fatty acid and stable isotope biomarkers. I apply a combination of observational and experimental approaches (both in the lab and in situ) to track the transfer of these biomarkers from algal diets to consumers.
by gway, last updated 6/26/15, sharing set to Public
This is example code for the Fatty Acid Source Tracking Algorithm in R (FASTAR) mixing model approach. This model was adapted from updated versions of MixSIR (Moore and Semmens 2008) to be used with fatty acids (FA) as variables [i.e., rather than stable isotopes (SI)] in the mixing model to identify proportional utilization of basal resources by invertebrate herbivore consumers. To date, the model has been applied to several field applications, a few of which are currently published/in press:
- Galloway et al. 2014. Diet-speciﬁc biomarkers show that high-quality phytoplankton fuels herbivorous zooplankton in large boreal lakes. Freshwater Biology 59:1902-1915. doi:10.1111/fwb.12394
- Galloway et al. 2014. Quantitative estimates of isopod resource utilization using a Bayesian fatty acid mixing model. Marine Ecology Progress Series 507:219-232. doi: 10.3354/meps10860
- Strandberg et al. 2015. Inferring heterogeneous phytoplankton composition with a fatty acid mixing model. Ecosphere 6:art16. doi: 10.1890/ES14-00382.1
- Galloway et al. 2015. A fatty acid based Bayesian approach for inferring diet in aquatic consumers. PLoS ONE 10(6):e0129723. doi: 10.1371/journal.pone.0129723
The FASTAR approach takes advantage of the fact that FA can differentiate algal groups at a resolution (e.g., Class level) that SI cannot. In addition, a FA based approach allows one to avoid an 'underdetermined' mixing problem by having many more variables than basal resources in a mixing model. These issues (and important details) are discussed in the first two papers cited above. A FA based mixing model approach for consumer dietary inference requires that the user build a resource library of FA signatures of consumers fed diverse basal resources. It is critical to account for diet-to-consumer trophic modification of FA by conducting experimental feeding trials with the consumers of interest. We have now done this for freshwater Cladocerans and marine isopods.
Here in this project, I am sharing the most up to date version of the FASTAR script and two dependent example files for both studies. Each analysis (i.e., isopods or cladocera) needs a producer and a consumer file. Please carefully refer to the comments in the script about the meaning and use of the files.
Log of changes: I will make changes to the files when we improve them (generally when we make the script more flexible for working with datasets with different dimensions). I will share older versions of the script with anyone who is interested, but I take them down from this site for the sake of simplicity (the updated files have not changed any of the original results they generated).
- 11-Sep-2014: I updated the general FASTAR script today, removing a line of code that provided the option of running multiple consumers at once in a given analysis, which is not what we did in the papers; the script used in the isopod paper ran each sample independently and compiled the results of these individual analyses at the site level.
- 12-Nov-2014: The new script (v4.1) will now adapt to whatever number of sources are in any producers file and allows for different numbers of 'groups' of consumers in the consumers file. It is run as a source file, and the script iteratively produces figures and .csv results files for each unique group identified in the consumer file (see example files). Each 'replicate' consumer is still shown in the figures. The heavy line in the distribution is the compiled 'group' solution and is the same as the results in the associated .csv. If you want a unique .csv for each consumer (i.e., the individual lines in the figures), simply set the group number = to the id number in the consumers file.
- 21-Jan-2015: I added a new folder below (FASTAR_v4_PLoSONE-revision) where the script and dependent files used in our PLoSONE revision manuscript can be accessed by the reviewers.
- 26-Jun-2015: I updated the (FASTAR_v4_PLoSONE) folder below so that it contains the exact dependent files and script now published and also available as a supplement on the PLoS ONE website.