SAN FRANCISCO (GenomeWeb) – Bisulfite sequencing has long been the gold standard for analyzing methylation despite shortcomings such as its high input DNA requirements and the fact that the chemical treatment damages the DNA.
Now, however, researchers from University of Oxford's Ludwig Institute for Cancer Research have developed a method for detecting at single-base resolution both 5-methyl cytosine and 5-hydroxymethyl cytosine that gets around those limitations of bisulfite conversion. The work was published on Monday in Nature Biotechnology, and the researchers aim to commercialize it and use it to detect epigenetic modifications on circulating tumor DNA.
Chuan He, professor at the University of Chicago and head of the Center of Excellence in Genomic Science, who was not involved with the study but has also developed epigenetic sequencing methods, said that the Ludwig team's method "has the potential to induce much less DNA degradation [than bisulfite conversion] and could be a really useful addition" to the field.
Chunxiao Song, a co-senior author of the study and assistant member of the Ludwig Institute for Cancer Research, said that he has been interested in using epigenetics in clinical cancer research, particularly methylation marks on cell-free DNA to help in cancer diagnosis and early detection.
"But, before we get to that point, we need the tools," he said. "Bisulfite sequencing is the gold standard and is very useful, but has always been unsatisfactory."
Bisulfite sequencing is indirect — by converting unmethylated cytosines into uracil, researchers must infer that the remaining unconverted cytosine bases are in fact methylated. In addition, Song said, the conversion results in an "unbalanced genome." Because the vast majority of cytosine bases are not methylated, converting those bases into thymine creates problems for the downstream mapping and analysis. Another disadvantage is that the chemical treatment itself is harsh and ends up degrading much of the DNA. To counter that, researchers use large amounts of input DNA, but for limited samples like cell-free DNA, that is not feasible.
These limitations of bisulfite sequencing spurred Song and the team to see whether they could develop a different method. To do this, they combined a known enzyme, TET, with novel chemistry.
The method, dubbed TET-assisted pyridine borane sequencing, or TAPS, also has an advantage in that it does not use harsh chemicals and so does not damage the DNA.
The researchers have filed a patent on the method and Song said that they are exploring options for commercializing it, including spinning out a company or licensing the technology to an existing company that would further develop it.
The method uses TET, a naturally occurring enzyme, to oxidize 5-mC and 5-hmC into 5-carboxylcytosine. Researchers, including Chicago's He, have previously made use of TET to develop epigenetic sequencing methods. He's group, for instance, used TET in combination with bisulfite sequencing to quantify 5-hmC.
The novel component of the Ludwig group's method came after converting 5-hmC and 5-mC into 5-caC. Through experimenting on 5-caC and screening chemicals that could react with 5-caC, the researchers found that borane-containing compounds could react with 5-caC, ultimately reducing it to a uracil derivative known as DHU. Because DHU is a uracil derivative, it is recognized by DNA and RNA polymerases as thymine.
After this conversion, sequencing proceeds as normal, after which the data is mapped back to the genome. And, "whenever there is a C-to-T mutation, that indicates the position" of 5-mC or 5-hmC, Song said.
In addition, the researchers also developed a way to isolate just one of the methylation marks. To do this, they used beta-glucosyltransferase to label 5-hmC, protecting it from TET oxidation and borane reduction. Only 5-mCs are converted to thymine and the 5-hmC positions can be deduced by subtraction.
To test the TAPS method, the researchers compared it with whole-genome bisulfite sequencing on genomic DNA from mouse embryonic stem cells. The team used 100 nanograms of input DNA for the TAPS method and 200 ng of input DNA for bisulfite sequencing. They also used spike-in controls to estimate the false-negative and false-positive rates of TAPS.
The TAPS method had a false-negative rate of between 2.7 percent and 3.5 percent and a false positive rate of 0.23 percent, which the researchers said was comparable to bisulfite sequencing. They cited a study that compared nine commercial bisulfite sequencing kits and found average false-negative and false-positive rates of 1.7 percent and 0.6 percent, respectively.
In addition, Song noted that the bioinformatics for TAPS is much simpler than for bisulfite sequencing. Bisulfite sequencing requires the data to be mapped to do different genomes, Song said, "so it takes a lot of computational resources." By contrast, the TAPS method can use standard mutation detection informatics. It's about three times faster than analysis for bisulfite sequencing, and Song noted it is less expensive. "We spent the same amount of money on both methods, but got twice as much useful data with TAPS compared to bisulfite sequencing," he said.
Song noted that the team plans to now use the method on cell-free DNA. The researchers have demonstrated proof of principle for this application and the first study is underway, Song said.
In addition, Song said that because TAPS does not damage the DNA, it could also be combined with long-read sequencing technology like Pacific Biosciences and Oxford Nanopore Technologies. He said that the team is currently testing the method on Oxford Nanopore's MinIon and is collaborating with other research groups to test it on PacBio's Sequel instrument.
Researchers have also developed methods to detect epigenetic marks directly on both PacBio's and Oxford Nanopore's instruments. On the PacBio instrument, epigenetic modifications can be detected by measuring changes in the polymerase kinetics — when the polymerase encounters a DNA modification as it is incorporating nucleotides, there is a measurable pause. Epigenetic modifications can be detected on the MinIon using bioinformatics that detect electrical current differences that result from a modified base translocating through the pore, as described in a Nature Methods study published in 2017.
Song said that one potential advantage of TAPS, at least in conjunction with PacBio sequencing, is that it would not require as deep coverage sequencing compared to the polymerase kinetics-based method.
He added that detecting methylation with long-read sequencing would be beneficial in understanding the role of methylation in genomic regions that are hard to access with short reads, like repetitive regions and centromeres.
Finally, the team is also interested in further developing TAPS to be useful at the single-cell level.
Song said that his lab plans to focus on using TAPS in oncology applications and especially for detecting methylation in cell-free DNA and is also interested in developing it for single-cell sequencing applications. However, he noted that TAPS could have much broader utility and could be used wherever bisulfite sequencing is traditionally used and even has the "potential to replace bisulfite sequencing in routine use."