This article was originally published Aug. 13.
A paper published last week that demonstrates the feasibility of the Helicos BioSciences technology for sequencing a human genome serves as an important validation of the system, but has also raised some questions about the relative cost — and ideal application of — the single-molecule sequencing technology.
Scientists from Stanford University reported in a letter in Nature Biotechnology last week that they have sequenced the genome of a Caucasian man — Stanford bioengineer and Helicos co-founder Stephen Quake — using a Helicos HeliScope single molecule sequencer and four data collection runs.
The paper marks the ninth published human genome sequenced on a next-generation sequencing platform, but the first to be sequenced with a single-molecule approach. According to Quake, the senior author on the paper, it represents "the first demonstration that you don't need a genome center to sequence a human genome."
In a statement, Quake, the primary inventor of the Helicos technology and chair of the company's scientific advisory board, said that the paper serves as proof that human genome sequencing "can now be done in one lab, with one machine, at a modest cost."
The team spent about four weeks and $48,000 on reagents to sequence the genome, Norma Neff, a research technical manager in Quake's lab and a co-author on the paper, told In Sequence sister publication GenomeWeb Daily News.
She said the main advantage of the HeliScope over other sequencing platforms is the high output and ease of library preparation, which does not require DNA amplification or ligation prior to sequencing. "If I had three [HeliScopes] I could do a genome a week," Neff said.
"Obviously this is sort of a landmark publication for Steve Quake and, because of that, it's also a landmark publication for Helicos," Patrice Milos, vice president and chief scientific officer for Helicos, told In Sequence. "We're really excited and thrilled."
Milos said Helicos has been working to create a machine that is easy to run and relies on simple upfront sample preparation — features that should make it feasible for a small group of individuals to do genome sequencing in independent labs rather than just in large genome sequencing centers.
"It really is exciting to think of how fast the technology has advanced," Milos said.
Prior to the Quake paper, the most significant published proof of concept for genome sequencing on the Helicos system was the publication in Science last year of the 6.4-kilobase M13 virus (see In Sequence April 8, 2009). Since then, the company has presented other genome sequencing projects, including of bacteria and small eukaryotes, at scientific conferences.
"What's clear about it is that they've done this with a new technology and it seems to work reasonably well," Chad Nusbaum, co-director of the Broad Institute's Genome Sequencing and Analysis program, told In Sequence.
Nusbaum, who was not involved in the study but has access to a Helicos instrument installed at the Broad, said the results presented in the paper look pretty good, especially since the researchers were dealing with fairly short (32 base), unpaired reads. "Considering that, I think they did very well."
Nusbaum was a bit more critical of the paper's insinuation that it takes "at least 35-40 machine runs with many instruments operating in parallel" and reagent costs of a quarter to a half a million dollars to sequence a human genome on other instruments, however. Those figures might have been true at one time, Nusbaum said, but not now.
"I'm very surprised that neither the reviewers nor the editors busted them on that," he said, noting that Illumina's genome sequencing service is currently priced at $48,000 — the same amount that the authors spent on reagents in the Quake paper.
[ pagebreak ]
Illumina announced its retail human genome sequencing service this June (see In Sequence June 16, 2009). The service relies on the company's Genome Analyzer II and offers about 30-fold genome coverage. Meanwhile, the personal genomics company Knome is offering whole-genome sequencing and interpretation for $99,500 and exome sequencing for $24,500, and Complete Genomics plans to offer a human-genome sequencing service for as little as $5,000.
In a paper published in Genome Research in June, ABI researchers reported that they had sequenced the genome of a Yoruban man to about 18 times coverage of the haploid genome. The cost of that effort: less than $30,000, according to the study's authors.
Representatives from Illumina were not available for comment before the time of publication. Roche/454 spokeswoman Katie Montgomery told In Sequence that Roche did not have any specific comments about the publication or its claims about genome sequencing costs at this time.
To sequence Quake's genome to about 28 times coverage, the researchers generated billions of reads that were between 24 and 70 bases long, with an average read length of 32 bases.
Because a small percentage of nucleotides remain unlabelled, Neff explained, there are some "dark" nucleotides that appear as deletions during the sequencing process. The team used Helicos software to assist in base calling and also corrected deletion errors by increasing their coverage of the genome, she said.
After mapping these reads to the human reference genome with a new software program called IndexDP, the researchers found that their reads covered 2.5 billion bases, or 90 percent of the reference.
They subsequently identified 2,805,471 SNPs and 752 copy number variants in the genome — results that they verified using selective Sanger re-sequencing, Illumina Human610-Quad SNP arrays, and comparisons with dbSNP and the Database of Genomic Variants.
The authors noted that there are still limitations to this and other sequencing technologies — including incomplete genome coverage and a lack of complete data on SNPs and structural variation.
"From a technophile standpoint, it's pretty cool," Kevin McKernan, senior director of SOLiD scientific operations for Life Technologies, told In Sequence. But, he said, by focusing on SNPs and large CNVs, the paper did not address the vast majority of the variation in the genome. That has him wondering why other types of variation such as small insertions and deletions were not analyzed. "Is it a fundamental limitation of the technology or is there more to come on this?" he mused.
Based on the data generated in the current paper, Quake and his colleagues argued that the HeliScope is a tool that could be used in "virtually any lab using a single commercial instrument" and could eventually "democratize access to the fruits of the genome revolution and may enable rapid and widespread adoption of individual genome sequencing in various scientific and medical contexts."
Though he said it's not entirely clear how Helicos sequencing costs compare with other platforms on the market right now, Nusbaum said he does believe it is cheaper on a per-read basis — savings that could make the instrument a reasonable investment for large sequencing centers.
But with a list price of $999,000, he questioned how feasible it would be for smaller labs to justify spending grant or startup dollars on a HeliScope. "To get into the game, that's a hard number to get over," Nusbaum said, explaining that unless a sequencing platform offers a significant advantage over its cheaper competitors, it could be difficult for small labs to raise the capital needed to buy the instrument.
As for immediate future applications of the technology, Stanford's Neff said she and her co-workers plan to continue using the HeliScope for human genome sequencing — particularly looking at cancer genomes.
For his part, though, Nusbaum said he's not sure whether whole-genome sequencing is the ideal application for the Helicos platform. He called the new paper a nice result and said he is excited about the technology in general, but said he thinks the technology is better suited to identifying copy number variations or applications such as chromatin-IP sequencing, where the accuracy of each base doesn't matter as much.
Nusbaum's team is currently testing several different applications for the HeliScope. For instance, at the Biology of Genomes meeting this spring, they reported on their work to find origins of replication in fission yeast genomes using the instrument (see In Sequence May 12, 2009).
"Helicos has a couple of really cool things about it," Nusbaum said, explaining that since it does not rely on DNA amplification, which may bias results, it provides a good picture of what's actually happening in the cell. In particular, he suspects it will be a good platform for transcript profiling in cells. "You're close to the natural state of the data," he said. "This has to be more accurate transcript profiling."
Despite the challenges that might remain for Helicos, Nusbaum said the company has at least one big advantage over its "third generation" single-molecule competitors such as Pacific Biosciences and Oxford Nanopore Technologies: "They're here," he said. "[Helicos is] going through these growing pains and testing pains that each of these [new platforms] will have to go through."