By Monica Heger
Researchers at Life Technologies recently demonstrated whole-methylome analysis of a bacterial genome using bisulfite sequencing on the SOLiD for the first time.
In a paper published last month in PLoS ONE, the Life Tech researchers described how they sequenced the methylome of E. coli using both an aqueous and an in-gel library construction process. The company plans to develop and commercialize a library construction kit for the aqueous method, In Sequence has learned.
Other researchers have developed techniques for methylome sequencing using the Illumina platform, but this study marks the first time it has been demonstrated on the SOLiD. The researchers said the key difference between methylome sequencing on the two platforms is that the Life Tech method uses SOLiD's color-coded two-base encoding scheme to identify all four bases whereas the Illumina approach is limited to three bases.
Life Technologies said it does not yet have a release date for the kit, nor has it made the cost of the kit public.
In the study, the Life Tech researchers started with 5 micrograms of bacterial DNA. The bisulfite conversion, which converts unmethylated cytosine into uracil while leaving the methylated cytosine intact, was done either in solution or on a polyacrylamide gel. They then amplified the libraries using emulsion PCR and sequenced them with 50-base fragment reads on the SOLiD.
Christina Chung, senior scientist at Life Technologies and lead author of the study, said that the two methods yielded comparable results. She said the main advantage of the gel-based construction is that it enables researchers to use less starting DNA than the aqueous approach. Since completing the study, Chung and her colleagues have begun using reagents from Life Tech's Invitrogen business that have allowed them to use even less starting material for that method than reported in the paper, she said.
Chung said the company only plans to commercialize the solution-based method, however, because it fits in better with other protocols for the SOLiD platform.
Chung said the main differentiator for the SOLiD technique is that methylome sequencing on the Illumina platform only allows for the analysis of three out of four bases because in the bisulfite conversion step the unmethylated cytosines are converted to thymines. Other researchers have gotten around this by using algorithms to figure out which thymines were originally cytosines.
SOLiD, on the other hand, uses a color-coded scheme, where one color corresponds to two bases. So even though the cytosines are converted to thymines, the four different colors are maintained throughout the process, Chung said.
"Since you use color codes and always look at two bases at a time, you can retain all four colors. So the advantage is that your genome is not as reduced," she said.
According to the authors, the technique may have advantages in applications such as methylome sequencing of cancer samples. It was "sensitive enough to identify partially methylated sites," which will give it an advantage in "applications of bisulfite sequencing in which copy number variation of a specific methylation motif is of biological importance," they wrote in the paper.
Chia-Lin Wei, a senior group leader at the Genome Institute of Singapore, has used Illumina to do methylome sequencing of human stem cell derivates and is currently testing methylome sequencing on the SOLiD.
She said that the two platforms so far seem comparable. She agreed with Chung that the color-coding scheme of the SOLiD allows the genome to maintain its complexity because there are still four colors, as opposed to three bases in the base-space scheme. This feature should allow for more accurate mapping, she said.
However, for more complex genomes with long strings of methylated cytosines, the SOLiD approach might lose its accuracy. In the current study, the researchers allowed for five mismatches in one read. More than four methylated cytosines in a row, said Wei, would have produced more than five color space mismatches, so the read would not have aligned correctly. That problem can be fixed by changing the algorithm to allow for more mismatches, she said. But the more mismatches you allow, the harder it is to correctly align the reads.
"You have a tradeoff," said Wei. "If you allow more mismatches, you will reduce mapping ability, but will detect more methylation. In the bacteria genome, it works well with five mismatches. It will be interesting to see, in a mammalian genome, what is the balance between mapping specificity versus detecting methylated CGs in a row."
In previous methylome sequencing studies using Illumina, the GIS group used single-end reads, but Wei said the paired-end reads would increase Illumina's accuracy. "The bottom line is that both platforms work well. The key is knowing their technical specifications so you can play around with the informatics mapping," she said.