Friday, July 25, 2014

Comparative Performance of Two Whole Genome Capture Methodologies on Ancient DNA Illumina Libraries

, , , , , , , ,

(Link) open access


1. The application of whole genome capture (WGC) methods to ancient DNA (aDNA) promises to increase the efficiency of ancient genome sequencing.

2. We compared the performance of two recently developed WGC methods in enriching human aDNA within Illumina libraries built using both double-stranded (DSL) and single-stranded (SSL) build protocols. Although both methods effectively enriched aDNA, one consistently produced marginally better results, giving us the opportunity to further explore the parameters influencing WGC experiments.

3. Our results suggest that bait length has an important influence on library enrichment. Moreover, we show that WGC biases against the shorter molecules that are enriched in SSL preparation protocols. Therefore application of WGC to such samples is not recommended without future optimization. Lastly, we document the effect of WGC on other features including clonality, GC composition and repetitive DNA content of captured libraries.

4. Our findings provide insights for researchers planning to perform WGC on aDNA, and suggest future tests and optimization to improve WGC efficiency.

Fig. 1. WGC preferentially retrieves longer fragments in sequencing libraries. The read length distribution of pre-capture and post-capture libraries is shown for (a) double-stranded libraries (DSL) and (b) single-stranded-libraries. In (a) the x axis is split in <90 bp and >90 bp to adjust the scale and better illustrate the higher concentration of short reads in the pre-capture libraries (pink line) and the bias observed against these in capture experiments (green and blue lines) where longer fragments are preferentially retrieved. (b) Illustrates that the relative gain of shorter fragments obtained by building a SSL, is lost by capturing these types of libraries. The plot shown in (c) depicts the bioanalyzer profile of the bait libraries revealing that for WISC a wider tail is observed for longer baits, which might explain the stronger bias in favor of longer fragments by this particular method.


By comparing the performance of WISC and MYbaits in enriching for endogenous human DNA in ancient DNA extracts, we have been able to pinpoint potential factors influencing the dynamics of WGC experiments. The assessment of the subtle differences between both approaches to in-WGC enables us to draw insights on two variables that may the affect capture efficiency – bait length distribution and hybridization time. Our data furthermore provides insights into the effect of blocking agents, and first insights into the performance of whole-genome enrichment methods on SSL.

Although the experimental design and parameters used in this study seem to suggest an apparent benefit of one of the methods over the other, we strongly caution that batch effects could be playing an important role under these settings, hence discourage such interpretation from our results. Likewise, it is worth considering that even though there is a certain convenience in using a pre-made kit (MYbaits), our observations point to specific factors that can be optimized in the in house method (WISC) namely bait length distribution and hybridization parameters. Knowing the relevance of such parameters in WGC, gives users the flexibility of customizing their capture experiment to match the particularities of each aDNA library (see below).

Role of bait length distribution on the efficiency of WGC

Bait length distributions (Fig. 1c) differ mainly in that WISC shows a wider range and longer bait lengths. This in principle could account for the marked retrieval of longer reads in the WISC compared with the MYbaits experiments (Figs. 1a-b).  Following this rationale, the higher success of the latter could be explained by its ability to better access a fraction of the sequencing library, specifically that with the smaller fragments, while this fraction remains inaccessible due to the higher concentration of longer baits used in WISC. An important consequence of this feature was the poor and even unsuccessful outcome of capturing SSL, which include a higher fraction of short fragments, with either method. At the same time, this limitation reveals an important area for future development in the context of WGC experiments.

Although there was a small, yet consistent, benefit in the MYbaits over the WISC, it would be rash to conclude that the MYbaits method always outperforms WISC.  These results were generated using a single batch of WISC bait versus a single batch of MYbaits. Given that (i) bait lengths will likely vary between batches as a result of initial template DNA fragmentation, and (ii) our hypothesis that bait length may play a key role in retaining shorter DNA fragments, we believe it is more than likely our results simply reflect the fact that in these batches tested the WISC bait were slightly longer than the MYbaits. Future studies that examine the role of bait length in capture success will be needed to further examine this hypothesis.

No comments:

Post a Comment

Comments have temporarily been turned off. Because I currently have a heavy workload, I do not feel that I can do an acceptable job as moderator. Thanks for your understanding.

Note: Only a member of this blog may post a comment.