Skip to main content
Premium Trial:

Request an Annual Quote

UW Team Adapts Transposase Library Prep for Ultra-Low-Input Whole-Genome Bisulfite Sequencing


By Molika Ashford

In a push to improve whole-genome methylation sequencing, University of Washington researchers who helped demonstrate a transposase-based library preparation method for DNA sequencing two years ago have now adapted the technique for bisulfite sequencing.

The researchers — Jay Shendure and Andrew Adey from UW's department of genome sciences — reported that their "tagmentation" method adapted for bisulfite sequencing required more than 100 times less starting material relative to current conventional methods and allowed the team to generate highly complex sequencing libraries from as little as 10 nanograms of input DNA and useful sequence from as little as one nanogram of DNA.

Shendure and Adey described the method in a paper published online in Genome Research last month. In the paper, the two describe testing the bisulfite sequencing library construction strategy, which they dubbed Tn5mC-seq, on a range of sample input sizes down to one nanogram, comparing it with a standard ligation-based approach.

"We're huge fans of this tagmentation [method.] The methylation protocol is really a natural extension," Shendure told In Sequence.

The scientists hope their adaptation of the transposase method will help overcome the need for large sample sizes in whole-genome bisulfite sequencing; something they say is a key limitation to whole-methylome research in many types of cells or samples.

After demonstrating the transposase library preparation method for shotgun DNA sequencing in 2010, Shendure's lab completely switched over to the method, sold by Epicentre, now Illumina (In Sequence 12/21/2010).

The original transposase method relies on a hyperactive version of the Tn5 transposase, which simultaneously fragments DNA and adds adaptors. The resulting products can then be subjected to PCR amplification and high-throughput sequencing. Epicentre introduced the first kits for the method several years ago, for use with the 454 and Illumina sequencing platforms.

Because bisulfite treatment yields single-stranded DNA and the target of the transposition reaction is double-stranded DNA, Shendure and Adey had to make several modifications so that the tagmentation reaction could take place before bisulfite treatment.

Adey explained that the researchers added an "oligo replacement strategy," that performs a gap-fill and ligates on methylated adaptors. This creates adaptor-flanked DNA, which can then be bisulfite-converted, Adey said.

Adey and Shendure said the adapted method is much easier than other approaches to whole-genome bisulfite sequencing, requiring fewer steps and reagents. "The ease of use is much better," Adey said.

Most importantly the method allowed the researchers to use much smaller samples than conventional methods require. In their report, the two researchers showed they could construct high-quality bisulfite sequencing libraries from as little as 10 nanograms of DNA from a lymphoblastoid cell line, with comparable results to libraries constructed using ligation chemistry with more than 100 times greater input.

"We sequenced to around 8.6-fold coverage just to show that we get generally the same methylation patterns, and it matched up really well with the standard library preparation method," Adey said.

The team also found that the tagmentation method could produce "useful sequence" — though not as high quality — from only one nanogram of DNA.

Shendure said the team's results show that the strategy could improve the practicality of whole-methylome sequencing. Because of the large sample requirements of ligation chemistry-based methods, researchers have previously turned to more targeted methylation sequencing methods in areas where DNA quantity is a bottleneck, like microdissected tissues or pathologies like cancer.

The efficiency of the method could also allow for preparation of whole genome bisulfite sequencing libraries from "poor-quality or degraded DNA samples," the authors wrote.

"Whether to do whole-genome [or targeted methylation sequencing] is a contemporary debate in the field, somewhat analogous to the exome-genome debate," Shendure said.

"We were interested in developing the technology to enable the field and the reason we focus on [the whole methylome] I think is that the longer-term trends are toward whole-genome bisulfite sequencing."

"For methylation, a number of papers have shown that gene body methylation and non-CpG island methylation [is important] … so we're still learning where the most relevant stuff is."

Joseph Ecker, a Salk Institute researcher who has worked with the standard ligation method Shendure and Adey used as a comparison to their tagmentation approach, told In Sequence in an e-mail that the transposase approach is an "important advance."

"Unlike most cases for genome sequencing where the amount of DNA is not limiting, a major challenge for whole-methylome sequencing is the often very low numbers of cells that investigators would like to interrogate," he wrote.

Ecker suggested that Shendure and Adey's method may make whole-methylome sequencing a more attractive option over reduced representation bisulfite approaches which have previously been the only bisulfite sequencing choice for researchers working with very small sample volumes.

He said he expects the approach to be "widely adopted." While he cited bias as a potential concern, he acknowledged that this issue was "not so bad" in Shendure's study.

According to Shendure and Adey, the same bias concerns raised in their initial work with the transposase method still hold for the methyl-seq adaptation. In their previous work, the two reported that the method was slightly more biased than others with regard to where the genome is fragmented, and that bias was greater for sequencing PCR amplicons, compared to genome sequencing.

"The bias shown with the bisulfite method mimics the bias that was shown with [our] previous DNA sequencing paper," Adey said.

"There is greater bias on the local level than [with] standard sequencing, but the effect is subtle and slight, and in our opinion not a major concern," said Shendure.

The researchers could not comment on the commercial context for their work and declined to detail any plans to market the adaptation on either their or Illumina's part.

Have topics you'd like to see covered in In Sequence? Contact the editor at mashford [at] genomeweb [.] com.