Final Report Summary - BEARMOLECULARDIET (A universal approach for analyzing the diet of omnivorous mammals based on DNA barcoding and next generation sequencing: application to the brown bear population in northern Italy)
The current brown bear (Ursus arctos) population in northern Italy is the result of a recent translocation, and is of great conservation interest because is one of the smallest and endangered bear population in Europe and it is central for the restoration of a bear population in the Alps (Swenson et al. 2000; Zibordi et al. 2010). Since the translocation, the population has been intensively monitored and has grown to >45 individuals expanding in the central Alps (De Barba et al. 2010; Groff et al. 2013). The modern Alpine landscape is highly impacted by human activities and is a mosaic of natural and anthropogenic habitats. Therefore understanding of the ecological processes in such a human-dominated environment is critical for the conservation of bears in this region. Diet studies are particularly important for endangered species or for population of conservation concern because accurate knowledge of animal diet is fundamental for evaluating the ecology of animals in their habitat. Further, as the distribution of food resources important to bears might be affected by habitat modifications due to climate change and human activities, monitoring of feeding habits constitutes a key aspect to predict how brown bears are influenced by, and could adjust to such changes in this region.
Diet studies of omnivorous animals and ecological understanding of trophic interactions between omnivores and their habitat is limited by the difficulty of accurately and efficiently determining the complex variety of food types animals have eaten in the field. The general goal of the project “A universal approach for analysing the diet of omnivorous mammals based on DNA barcoding and next-generation sequencing: application to the brown bear population in northern Italy” was to develop a universal and standardized tool for comprehensive molecular non-invasive diet analysis of omnivorous animals, to validate it using the brown bear population in northern Italy, and finally to evaluate its use for understanding brown bear ecology in this region highly impacted by human presence.
We developed a method based on DNA metabarcoding multiplexing and next-generation sequencing (NGS) to uncover different taxonomic groups of organisms from a complex environmental mixture in a single experiment, through the co-amplification of three universal DNA metabarcoding markers, targeting the plant (P6 loop of the chloroplast trnL (UAA), Taberlet et al. (2007)), vertebrate (V5 loop of the mitochondrial 12S gene, Riaz et al. (2011)), and invertebrate components (short fragment of the mitochondrial 16S gene, Taberlet unpublished) of an omnivorous diet. To increase taxonomic resolution for plants of the families Asteraceae, Cyperaceae, Poaceae and Rosaceae we complemented the information of the trnL with additional markers (short fragments of the internal transcribed spacer region 1 (Baamrane et al. 2012, Taberlet unpublished) and region 2 (developed for this study) of nuclear ribosomal DNA.
One aspect of diet studies based on DNA metabarcoding that requires further development, especially when complex food habits are under focus, relates to the need of validation of the quality of the final data produced. Infect, NGS technologies generate enormous numbers of reads for any DNA molecule amplified by PCR, which can result in the inclusions of erroneous sequences. We established a standard protocol of sequence data filtering based on the use of internal controls, PCR replication, and patterns in sequence distribution data across samples to guide and objectively justify the selection of filtering parameters in post-sequencing data analysis, and to ensure and allow evaluation of the accuracy of the final results.
We validated our DNA metabarcoding approach and data filtering protocol using 91 brown bear feces collected in northern Italy during a genetic monitoring program for the brown bears in this region (De Barba et al. 2010). Using the multiplexing strategy outlined above, we significantly simplified the experimental procedure and accurately, concurrently identified different food remains corresponding to the targeted taxonomic groups, with ≥60% of taxa of all diet components identified to genus/species level. The systematic application of internal controls and replication was a useful and simple way to evaluate the performance of our experimental procedure, standardize the selection of sequence filtering parameters for each marker data, and validate the accuracy of the results. Our general approach can be adapted to the analysis of dietary samples of various omnivorous species in different ecosystems, using both non-invasive (i.e. feces, pellets) and invasive samples (i.e. stomach contents), for studies entailing large-scale population level diet assessments through cost effective screening of multiple DNA metabarcodes, and the detection of fine dietary variation among samples or individuals and of rare food items. We expect this method to be of particular utility in a number of conservation and ecological applications for omnivorous animals including monitoring diet shifts as a consequence of natural habitat modifications under climate change and/or increased human activities, dietary studies of elusive and endangered species, resource-partitioning assessments among sympatric species, and detecting levels of exploitation of endangered species or anthropogenic food. In addition, our general multiplexing strategy can be applied to other biodiversity DNA metabarcoding based studies for the analysis of various types of complex environmental mixtures other than dietary samples.
We used the developed laboratory and data analysis protocol to examine the diet of the brown bears in northern Italy at the population and individual level. Three-hundreds-seventy brown bear samples collected in northern Italy in 2002-2009 and previously genotyped for individual and sex identification were used for the analysis. We determined the composition of 344 of the samples analyzed. Plants were found in 316 samples, vertebrates in 86, and invertebrates in 144. Among samples for which composition was determined, most were either entirely composed by plants (42%) or by a combination of plants and invertebrates (29%). The plant component was more diversified (57 different taxa), including a greater number of species belonging to different families, compared to the vertebrate (8 taxa) and the invertebrate (15 taxa) components. While these data show a highly diversified feeding regime reflecting the omnivore nature of brown bears, they also offer insights into the feeding ecology of this population in the study area and on the potential impact of human presence and activities. Among plants, species of the Asteraceae, Rosaceae and Fagaceae families had the highest frequency of occurrence in the samples reflecting a greater utilization by bears and the importance of seasonal food like Rubus idaeus and Fagus Sylvatica. On the other hand, the detection of species of genus Prunus, Malus and Pyrus in the Rosaceae family revealed use of plants commonly cultivated in the surrounding of the villages. Ants of the genus Formica and Lasius were most frequently found among invertebrates, and genus Ovis and Bos among vertebrates. The relatively high frequency of Bos in the samples can be most likely explained with diffuse farming in the Alpine region, and consequent availability of cattle carcasses for bears to feed upon.
Population level patterns of diet composition observed in the samples analyzed are concordant with microscopic examination of feces that revealed a greater occurrence and diversity of plants in the diet of brown bears in northern Italy compared to other dietary components. The two analyses differed in the proportion and taxonomic resolution of vertebrates identification, specifically mammals, with the DNA metabarcoding approach resulting in detection of vertebrates in a greater number of samples and at the genus/species level, and in the taxonomic resolution of invertebrates, with microscopic examination resulting in most species level identifications. This suggests a complementary application of the two techniques.
We combined diet profiles obtained for each sample with information on single individuals from which samples were collected. Our preliminary results show a greater consumption of arthropods for females (68% of samples containing invertebrates) and of vertebrates for males (56% of samples containing vertebrates). We also found a greater occurrence of cultivated plants, i.e. genus Prunus, Malus and Pyrus, among male samples (57% of samples containing these genera).
At the light of our experimental project, the DNA metabarcoding approach for omnivorous diet analysis proved to have enormous potentials to study the diet of brown bears in the Italian Alps. Its efficient application makes it particularly suited as a tool for effective management programs through increased understanding of the ecology of such large mammal in a human dominated environment.
References
Baamrane MAA, Shehzad W, Ouhammou A, et al. (2012) Assessment of the food habits of the Moroccan dorcas gazelle in M'Sabih Talaa, West central Morocco, using the trnL approach. PLoS ONE 7.
De Barba M, Waits LP, Garton EO, et al. (2010) The power of genetic monitoring for studying demography, ecology and genetics of a reintroduced brown bear population. Molecular Ecology 19, 3938-3951.
Groff C, Bragalanti N, Rizzoli R, Zanghellini P (2013) Rapporto Orso 2012 Servizio Foreste e fauna della Provincia Autonoma di Trento, Trento.
Riaz T, Shehzad W, Viari A, et al. (2011) ecoPrimers: inference of new DNA barcode markers from whole genome sequence analysis. Nucleic Acids Research 39, e145-e145.
Swenson JE, Gerstl N, Dahle B, Zedrosser A (2000) Action plan for the conservation of the Brown Bear (Ursus arctos) in Europe Council of Europe.
Taberlet P, Coissac E, Pompanon F, et al. (2007) Power and limitations of the chloroplast trnL (UAA) intron for plant DNA barcoding. Nucleic Acids Research 35, e14.
Zibordi F, Mustoni A, Viviani V, Liccioli S, Stefani G (2010) L'impegno del Parco per l'orso: il progetto Life Ursus Manfrini, Rovereto, Strembo, Italy.
−
Diet studies of omnivorous animals and ecological understanding of trophic interactions between omnivores and their habitat is limited by the difficulty of accurately and efficiently determining the complex variety of food types animals have eaten in the field. The general goal of the project “A universal approach for analysing the diet of omnivorous mammals based on DNA barcoding and next-generation sequencing: application to the brown bear population in northern Italy” was to develop a universal and standardized tool for comprehensive molecular non-invasive diet analysis of omnivorous animals, to validate it using the brown bear population in northern Italy, and finally to evaluate its use for understanding brown bear ecology in this region highly impacted by human presence.
We developed a method based on DNA metabarcoding multiplexing and next-generation sequencing (NGS) to uncover different taxonomic groups of organisms from a complex environmental mixture in a single experiment, through the co-amplification of three universal DNA metabarcoding markers, targeting the plant (P6 loop of the chloroplast trnL (UAA), Taberlet et al. (2007)), vertebrate (V5 loop of the mitochondrial 12S gene, Riaz et al. (2011)), and invertebrate components (short fragment of the mitochondrial 16S gene, Taberlet unpublished) of an omnivorous diet. To increase taxonomic resolution for plants of the families Asteraceae, Cyperaceae, Poaceae and Rosaceae we complemented the information of the trnL with additional markers (short fragments of the internal transcribed spacer region 1 (Baamrane et al. 2012, Taberlet unpublished) and region 2 (developed for this study) of nuclear ribosomal DNA.
One aspect of diet studies based on DNA metabarcoding that requires further development, especially when complex food habits are under focus, relates to the need of validation of the quality of the final data produced. Infect, NGS technologies generate enormous numbers of reads for any DNA molecule amplified by PCR, which can result in the inclusions of erroneous sequences. We established a standard protocol of sequence data filtering based on the use of internal controls, PCR replication, and patterns in sequence distribution data across samples to guide and objectively justify the selection of filtering parameters in post-sequencing data analysis, and to ensure and allow evaluation of the accuracy of the final results.
We validated our DNA metabarcoding approach and data filtering protocol using 91 brown bear feces collected in northern Italy during a genetic monitoring program for the brown bears in this region (De Barba et al. 2010). Using the multiplexing strategy outlined above, we significantly simplified the experimental procedure and accurately, concurrently identified different food remains corresponding to the targeted taxonomic groups, with ≥60% of taxa of all diet components identified to genus/species level. The systematic application of internal controls and replication was a useful and simple way to evaluate the performance of our experimental procedure, standardize the selection of sequence filtering parameters for each marker data, and validate the accuracy of the results. Our general approach can be adapted to the analysis of dietary samples of various omnivorous species in different ecosystems, using both non-invasive (i.e. feces, pellets) and invasive samples (i.e. stomach contents), for studies entailing large-scale population level diet assessments through cost effective screening of multiple DNA metabarcodes, and the detection of fine dietary variation among samples or individuals and of rare food items. We expect this method to be of particular utility in a number of conservation and ecological applications for omnivorous animals including monitoring diet shifts as a consequence of natural habitat modifications under climate change and/or increased human activities, dietary studies of elusive and endangered species, resource-partitioning assessments among sympatric species, and detecting levels of exploitation of endangered species or anthropogenic food. In addition, our general multiplexing strategy can be applied to other biodiversity DNA metabarcoding based studies for the analysis of various types of complex environmental mixtures other than dietary samples.
We used the developed laboratory and data analysis protocol to examine the diet of the brown bears in northern Italy at the population and individual level. Three-hundreds-seventy brown bear samples collected in northern Italy in 2002-2009 and previously genotyped for individual and sex identification were used for the analysis. We determined the composition of 344 of the samples analyzed. Plants were found in 316 samples, vertebrates in 86, and invertebrates in 144. Among samples for which composition was determined, most were either entirely composed by plants (42%) or by a combination of plants and invertebrates (29%). The plant component was more diversified (57 different taxa), including a greater number of species belonging to different families, compared to the vertebrate (8 taxa) and the invertebrate (15 taxa) components. While these data show a highly diversified feeding regime reflecting the omnivore nature of brown bears, they also offer insights into the feeding ecology of this population in the study area and on the potential impact of human presence and activities. Among plants, species of the Asteraceae, Rosaceae and Fagaceae families had the highest frequency of occurrence in the samples reflecting a greater utilization by bears and the importance of seasonal food like Rubus idaeus and Fagus Sylvatica. On the other hand, the detection of species of genus Prunus, Malus and Pyrus in the Rosaceae family revealed use of plants commonly cultivated in the surrounding of the villages. Ants of the genus Formica and Lasius were most frequently found among invertebrates, and genus Ovis and Bos among vertebrates. The relatively high frequency of Bos in the samples can be most likely explained with diffuse farming in the Alpine region, and consequent availability of cattle carcasses for bears to feed upon.
Population level patterns of diet composition observed in the samples analyzed are concordant with microscopic examination of feces that revealed a greater occurrence and diversity of plants in the diet of brown bears in northern Italy compared to other dietary components. The two analyses differed in the proportion and taxonomic resolution of vertebrates identification, specifically mammals, with the DNA metabarcoding approach resulting in detection of vertebrates in a greater number of samples and at the genus/species level, and in the taxonomic resolution of invertebrates, with microscopic examination resulting in most species level identifications. This suggests a complementary application of the two techniques.
We combined diet profiles obtained for each sample with information on single individuals from which samples were collected. Our preliminary results show a greater consumption of arthropods for females (68% of samples containing invertebrates) and of vertebrates for males (56% of samples containing vertebrates). We also found a greater occurrence of cultivated plants, i.e. genus Prunus, Malus and Pyrus, among male samples (57% of samples containing these genera).
At the light of our experimental project, the DNA metabarcoding approach for omnivorous diet analysis proved to have enormous potentials to study the diet of brown bears in the Italian Alps. Its efficient application makes it particularly suited as a tool for effective management programs through increased understanding of the ecology of such large mammal in a human dominated environment.
References
Baamrane MAA, Shehzad W, Ouhammou A, et al. (2012) Assessment of the food habits of the Moroccan dorcas gazelle in M'Sabih Talaa, West central Morocco, using the trnL approach. PLoS ONE 7.
De Barba M, Waits LP, Garton EO, et al. (2010) The power of genetic monitoring for studying demography, ecology and genetics of a reintroduced brown bear population. Molecular Ecology 19, 3938-3951.
Groff C, Bragalanti N, Rizzoli R, Zanghellini P (2013) Rapporto Orso 2012 Servizio Foreste e fauna della Provincia Autonoma di Trento, Trento.
Riaz T, Shehzad W, Viari A, et al. (2011) ecoPrimers: inference of new DNA barcode markers from whole genome sequence analysis. Nucleic Acids Research 39, e145-e145.
Swenson JE, Gerstl N, Dahle B, Zedrosser A (2000) Action plan for the conservation of the Brown Bear (Ursus arctos) in Europe Council of Europe.
Taberlet P, Coissac E, Pompanon F, et al. (2007) Power and limitations of the chloroplast trnL (UAA) intron for plant DNA barcoding. Nucleic Acids Research 35, e14.
Zibordi F, Mustoni A, Viviani V, Liccioli S, Stefani G (2010) L'impegno del Parco per l'orso: il progetto Life Ursus Manfrini, Rovereto, Strembo, Italy.
−