Some shoppers like to browse several stores, whereas others follow the Amazon approach and want to just visit one place to buy everything on their list. The lab of Shankar Balasubramanian’s (University of Cambridge) must be in that second group because they developed a sequencing method that is a true one-stop extravaganza. It rings up genetic and epigenetic reads down to the single-base level and can even be done with tiny amounts of DNA, such as cell-free DNA (cfDNA), which is useful for prenatal and cancer research.
In the past, it’s been a toss-up between getting detailed genetic or epigenetic sequences from short-read sequencing. Some techniques, such as whole-genome bisulfite sequencing (WGBS) or enzymatic-methyl sequencing (EM-seq), can tell unmodified Cs from modified Cs, but these approaches mask real C-to-T genetic mutations because they convert the unmodified C to a U, which is then read as a T. And, without separate workflows, they can’t resolve 5-methylcytosine (5mC) from 5-hydroxymethylcytosine (5hmC).
So, Balasubramanian’s team put everything in one convenient workflow for the time-strapped shopper. Here’s how the method known as five-letter seq reads five different bases (the four usual suspects, plus 5mC):
- Fragment DNA and ligate it to hairpin adapters at both endsSplit open the structure and synthesize complimentary strands for each of the sample strands
- Protect 5mCs by transforming them to 5hmCs, or other mods like 5fC, with oxidation, and protect 5hmCs by glycosylation, creating different information states
- Strands are opened up and unmodified Cs are deaminated
- Because the products are no longer complementary, they can undergo PCR and sequencing
- A little computational know-how resolves the different states in the data to one of five “letters”
Six-letter seq, which goes the extra mile and differentiates 5mCs from 5hmCs, works in a similar way. The first steps are the same, but methylation at the 5mCs gets copied to the copy strand, and 5hmC is glycosylated to inhibit this copying. Then unmodified Cs get deaminated, and so on.
The team used this highly efficient all-in-one strategy with beta-lymphoblast human genomic DNA and cfDNA from a real colon cancer patient. The method was highly accurate, worked with very limited sample amounts, and is compatible with commercially available sequencing platforms.
Get everything you need all in one convenient place at Nature Biotechnology, February 2023.