Ultra-deep sequencing

Next generation sequencing in biodiversity research usually involves DNA extraction of bulk samples, PCR amplification of targeted markers and subsequent massive parallel sequencing. Those bulk samples can contain entire organisms or come from a single environmental sample containing degraded DNA. A drawback in this procedure is the PCR amplification step that is applied to acquire a sufficient number of barcode sequences. Primers are designed to amplify a large range of taxa present in the bulk sample but those are rarely universal. Although much effort has been made to increase the universailty and efficiency of primers it is difficult to predict their performance when an investigated fauna is largely unknown.

A group of researchers from the Beijing Genomics Institute have developed a new pipeline that is independent of PCR amplifications while still enabling DNA Barcode identification. They used ultra-deep sequencing to detect minute traces of mitochondrial sequences directly from genomic DNA mixtures.

For this proof of concept study 73 insects were collected, preserved and visually identified. Once complete (as far as possible) DNA from the legs of each creature was analyzed using DNA barcoding (with standard Sanger sequencing) and compared with data from the Barcode of Life Data Systems. Subsequently, after mitochondrial enrichment using differential centrifugation, total genomic DNA was extracted from a homogenized mixture of the remainder of the insects.

The group employed two different approaches to assign species-level identity: 

(1) By utilizing a DNA Barcode reference library for the investigated fauna. Sequence reads were mapped to the reference sequences following defined criteria.

(2) In absence of a reference library, a de novo assembly of sequence reads into mitochondrial gene fragments, especially the COI Barcode region, was employed, followed by gene annotation, to ensure accurate detection of taxa from the mixed bulk sample.

Both experiments worked quite well which I find very exciting. The team was able to assess biodiversity at the species level regardless of whether a species can be actually named. Their work also showed that it is possible to calculate the total quantity of mitochondrial DNA present which would allow to extrapolate the ratios and physical mass of each species in the mix.

Of course this proof of concept needs to be transferred into 'real-world' applications but once working it would represent a big step forward for biodiversity-related surveillance efforts, such as in biomonitoring programs.