I was at the Science for the Environment conference in Aarhus (Denmark) last week, which focussed 3 days on new trends and developments in the biodiversity monitoring community. It was really cool to see how remote drones are now used in monitoring, especially for plants. Also, creating citizen science projects for data processing has become really easy with projects like galaxyzoo. So, whats new for DNA based monitoring?
DNA Metabarcoding vs Metagenomics
There are currently two sequencing approaches to DNA based monitoring of animals: Targeted amplicon sequencing (metabarcoding) and direct PCR free sequencing (metagenomics). Both methods are useful for different research questions and monitoring tasks, but I was so far always in favour for DNA metabarcoding, as metagenomics required more sequencing effort and lacks mitochondrial reference genomes. However I discussed the perspectives of metagenomics with Kat Bruce from NatureMetrics and actually I’m really exited about the potential of this approach.
A metagnomics approach, potentially provides more taxonomic resolution, as the whole nuclear + mitochondrial genome is sequenced, given enough sequencing depth. However, for monitoring of animals COI metabarcoding has sufficient taxonomic resolution as already 150 bp of the Folmer region are sufficient to identify <90% of the taxa on species level (Meusnier et al. 2008).
DNA metabarcoding relies on amplification of a specific barcoding gene. However, in PCR not all taxa are amplified equally well, due to mismatches between Primer and DNA template. This leads to differences in sequence abundance of up to 4 magnitudes between different taxa (Elbrecht & Leese 2015). Unfortunately, this means that it becomes very difficult to estimate taxa abundance with PCR based methods. Additionally some taxa might not be amplified at all, and thus missed completely with the PCR based monitoring.
You can learn more about primer bias in DNA metabarcoding of macro invertebrates in my talk from the iBOL conference in Canada 2015 in the video below.
A metagenomics approach, or more specific a mitogenomics approach, potentially solves a lot of these issues as it does not rely on PCR amplification. By directly shotgun sequencing the DNA of a bulk sample, one would expect to detect all taxa in the sample with the read number being proportional taxa biomass. First studies exploring this PCR free approach show promising sequence abundance / taxon biomass relationships and good taxa detection rates (Gómez‐Rodríguez et al. 2015, Tang et al. 2015). Metagenomics has a lot of potential for biomonitoring, however methodological biases and potential shortcomings of this approach have to be further evaluated and verified.
While assessment based on presence and absence of OTUs is possible, and probably the only way to go if the decline of taxonomic experts continues, it would be really great if we could assign species level taxonomy to sequences. For COI DNA metatabarcoding this is often possible, as databases like BOLD now contain over 5 million COI reference sequences. Getting an accurate taxa list from an ecosystem allows for detailed analysis, and the potential to answer many ecological questions that go beyond the basic monitoring.
While metagenomics can utilise the existing COI barcoding databases as well, it would be desirable to use complete mitochondrial genomes as references, to map and use more reads in the analysis. With high throughput sequencing, it has now become fairly easy to generate mitochondrial references. Unfortunately, many researchers are still neglecting the importance of generating mitochondrial genomes, and “just don’t care”. But even DNA metabarcoding would profit from mitochondrial genomes, enabling the evaluation of alternative markers like 16S which might show less primer bias. I would love to see more colleges embracing the “boring” sequencing of mitochondrial genomes. With comprehensive mitochondrial reference databases we can unlock the full potential of metagenomics.
Currently a metagenomics approach requires ~100x sequencing depth in comparison to DNA metabaroding, as metagenomics can only fall back on the COI reference databases with only few mitochondrial genomes publicly available. Once enough mitochondrial reference genomes are available, more of the metagenomic data can be used for analysis, allowing for lower sequencing depth and thus driving down sequencing cost.
Additionally, Techniques for enriching for mitochondrial DNA are being developed, increasing the proportion of mitochondrial reads in the library to up over 40% (Liu et al. 2015). When combining a good reference database, containing full mitochondrial genomes with techniques for enriching for mitochondrial reads, the sequencing cost of DNA metabarcoding and metagenomics could be similar. However, currently DNA metabarcoding is the cheaper option.
If you would have ask me a few weeks ago, I would have spoken in favour of DNA metabarcoding. However, recent papers indicate that the issues preventing a metagenomics approach being useful for biomonitoring of animals might be solved. With (hopefully) growing mitochondrial reference databases and mitochondrial enrichment, metagenomics can fix some of the shortcomings of DNA metabarcoding, delivering more complete taxa lists at possibly the same cost. I’m still not convinced that metagenomics will be able to reliably estimate abundance, but it might have more potential than a PCR based approach.
So which approach is the way to go for biomonitoring? I think both approaches have potential, and currently it’s really difficult to decide wich one will be more popular. Today, DNA metabarcoding has the edge, due to good COI reference databases and ~10-100x less sequencing cost. However, metagenomis is back in the game, promising more complete taxa lists and maybe even biomass estimates maybe even for the same price as DNA metabacrcoding. Let’s be open to both approaches, and let’s make some mitogenomes for our favorite taxa!
Elbrecht & Leese (2015). Can DNA-Based Ecosystem Assessments Quantify Species Abundance? Testing Primer Bias and Biomass—Sequence Relationships with an Innovative Metabarcoding Protocol. PloS One
Gómez‐Rodríguez, Crampton‐Platt, Timmermans, Baselga & Vogler (2015). Validating the power of mitochondrial metagenomics for community ecology and phylogenetics of complex assemblages. Methods in Ecology and Evolution
Liu, Wang, Xie, Tan, Li, Su, Zhang, Misof, Kjer, Tang, Niehuis, Jiang & Zhou (2015). Mitochondrial capture enriches mito-DNA 100 folds enabling PCR-free mitogenomics biodiversity analysis. Molecular Ecology Resources
Meusnier, Singer, Landry, Hickey, Hebert & Hajibabaei (2008). A universal DNA mini-barcode for biodiversity analysis. BMC Genomics
Tang, Hardman, Ji, Meng, Liu, Tan, Yang, Moss, Wang, Yang, Bruce, Nevard, Potts, Zhou & Yu (2015), High-throughput monitoring of wild bee diversity and abundance via mitogenomics. Methods in Ecology and Evolution