A rising star in the epigenetics field, Dr. Kun Zhang joined the UCSD faculty in 2007 after a post-doc with George Church at Harvard. Engaging in a little friendly competition with his former mentor, Zhang published a technique for targeted bisulfite sequencing using padlock probes (Nat. Biotechnol. 2009, 27, 353) at the same time the Church lab published a similar method (Nat. Biotechnol. 2009, 27, 361). In Zhang’s study, his group designed a set of ~30,000 probes to capture (in a single test tube) ~66,000 bisulfite-treated CpG sites on three human chromosomes. The captured DNA was amplified and analyzed by deep sequencing.
Funded by an NIH Roadmap Epigenomics technology development grant, Zhang is busy scaling up the method to tackle the entire methylome. EpiGenie caught up with him to discuss his new method and future plans.
Kun Zhang Interview
EpiGenie: The cost of whole-genome bisulfite sequencing, at least for human DNA, is still prohibitively high. How does your padlock-probes-based method for DNA methylation analysis help bridge the gap between genome coverage and affordability?
Zhang: The best method would be to do whole-genome bisulfite sequencing because then you can get an unbiased and most accurate characterization of DNA methylation across an entire genome. There are people doing that now with human DNA samples, but this is pretty expensive. It takes a lot of money and a lot of effort. The reason we developed our method is because we have been working at targeted sequencing of the human genome (actually, exon sequencing), and we developed a padlock-based method to do that. Then, we realized that that method can be translated to targeted bisulfite sequencing because if you have some reasonable guess at the genome region you want to look at, why do you want to spend your sequencing dollar on 99% of the other regions that you’re not interested in at all? So basically, our method allows you to select any kind of target you want on a really large scale and just focus your sequencing dollar on those targets and sequence really deep. Using our method basically allows you to retrieve information about the DNA methylome a lot more efficiently than other methods.
EpiGenie: What are some interesting findings you’ve obtained using this method?
Zhang: One really striking observation was that only 7% of the CpG islands showed methylation differences between two very different cell types. Most of the methylome didn’t change at all. Also, we found that for active genes, the gene body tends to be methylated. There have been one or two reports in that direction, but previously the common thinking was that methylation is a silencing mark. It turns out that for a gene to be activated, the promoter has to be unmethylated but the gene body needs to be more methylated.
EpiGenie: To some of us, designing a set of 30,000 padlock probes to target 66,000 CpG sites seems almost like a Herculean task. Are most labs going to be able to easily select CpG targets and design oligonucleotide padlock probes?
Zhang: We’ve provided a method that allows you to capture any target you want, and it depends on the investigators what their interests are. It’s their responsibility to come up with the list of CpG targets that is most relevant to their study. We just did a proof of concept saying that we can capture every single CpG island on three chromosomes, but our method is not limited to CpG islands.
We actually released software with our paper that is freely available. So anybody can design probes to any target they want and just synthesize the DNA, and they can capture any target. So the average user with a list of 5,000 or 10,000 genes can just copy and paste and click a button, and the software will spit out all of the oligo sequences, which can be uploaded to Agilent’s website for synthesis. The software we released in the paper is sort of version 1.0. It doesn’t have a nice interface, so you need some computer skills to run it. But now we’re working on version 2.0 where we put a very nice interface. We’re trying to make it easy for other people. It is expensive to synthesize the oligos. To bring this method to the average user, we really need to figure out a way to commercialize the method so that the average user doesn’t have to deal with designing or synthesizing oligos. Ideally, they could just go to a website, pick the gene they want, and order the padlock probes.
EpiGenie: What directions are you now taking with the padlock probes method?
Zhang: We are expanding this method to a variety of samples-cancer samples, various disease samples. Also, we’re funded by the Roadmap Epigenomics program to develop a whole-genome process. So in the paper we only show two chromosomes, but now we are scaling up to the entire genome using the same approach. Right now, we have a list of targets we’re interested in, and then we just design probes to cover those targets across the entire genome. You can just go to the genome annotation, and there are all kinds of functional elements, and basically we can just design probes to any single locus. So this is moving very fast.
EpiGenie: Speaking of the Roadmap Epigenomics program, how do you think this program will advance the field of epigenomics and help overcome some of the limitations with the current analytical methods?
Zhang: I think the real value of this program will be in the big mapping centers that are going to map all kinds of elements in the genome. Because right now, if you’re going to study whatever epigenetic mark is related to your disease, you don’t really have a very clear idea about where you should be looking. Those mapping centers are going to look at all of the epigenetic marks–DNA methylation, histone modifications, and any other kind of epigenetic modification that you can imagine, in many different cell lines–and they’re going to collate their efforts and publish this reference map so that we know where to look for anything we’re interested in. So that there will be a very, very highly valuable resource.
Those mapping centers are just using existing methods to generate data. We’re funded by the Roadmap Epigenomics technology development program, and we’re developing a newer-generation method that will be more efficient. When you have a map, you need a car to quickly get to whatever place you want to go. We’re developing a method to allow investigators to take advantage of those reference maps. So I think there’s a lot of synergy.