By Julia Karow
This story was originally published on January 14, 2010.
Illumina's planned increase in output for its Genome Analyzer to 95 gigabases per run has been delayed by about three months due to a reagent problem that the company disclosed last fall, according to the firm's CEO. The company now expects to reach that commercial milestone in the second quarter.
Recent orders for the instrument have been strong, CEO Jay Flatley said at the JP Morgan Healthcare Conference in San Francisco last week. He noted that during the fourth quarter, the company received, for example, 17 multi-system orders for a total of 61 instruments.
Flatley made his remarks during a presentation in which he also introduced a new sequencing instrument, HiSeq 2000, which uses the same sequencing chemistry as the GA but has a higher throughput (see In Sequence 1/12/2010). The company will continue to sell and support the GA.
Illumina planned to launch new reagent kits for the GA last week, he said, that will enable customers to generate 50 gigabases of sequence data per run. Several early-access customers have already been testing these kits, and these "are now getting routinely in excess of 55 [gigabases] per run, and a number of them have gotten over 60 [gigabases]," he said.
The 50-gigabase reagent kits will enable customers to sequence a human genome at a reagent cost of about $20,000, he said, while the HiSeq 2000 will allow them to sequence a human genome for half that price.
Originally, the company had planned to increase the output of the GA to 95 gigabases per run by the end of 2009, but due to a problem with defective paired-end sequencing reagent kits last fall (see In Sequence 11/3/2009), that development program was delayed by about three months, Flatley said, because "all our machines internally were occupied resolving that reagent issue."
He said Illumina has now had internal runs generating more than 95 gigabases of data, and these improvements will "go into early access this month and be fully commercial, we believe, in Q2."
The problem with the defective reagent kits, which reduced revenues in the fourth quarter by about $4 million to $5 million, was "fully resolved" last year, according to Flatley.
In addition to improving the output of the GA, and developing the HiSeq 2000, Illumina has been working on facilitating the front-end sample prep and back-end data analysis for its sequencers, in particular with regard to the high throughput of the HiSeq 2000.
In the second quarter, Flatley said, Illumina will start shipping new sample-prep kits that will make library construction for the systems "vastly simpler."
These kits will be automation-friendly, reduce the hands-on time required, cut the number of tubes needed in half, and reduce the clean-up and gel steps "very significantly," he said. Customers can use the kits with 96-well plates to prepare 96 samples simultaneously, and will be able to prepare a library about 12 times faster than previously.
Flatley said that the company has also unified protocols across different sequencing applications • such as DNA sequencing, expression analysis, and methylation analysis • which had "diverged a bit" in the past. Though the protocols will still differ, "every common step uses exactly the same methodology and the same reagents" now.
[ pagebreak ]
Illumina has also streamlined the data coming off its sequencers, he said. For example, it has "radically improved" the processing of image intensities from the instruments in real time, which is now possible on a desktop computer. Each read generation cycle now produces only 100 gigabytes of data instead of 14 terabytes in the past. For alignment and mapping, the file size has decreased to 600 gigabytes from 3 terabytes. A full human genome, "once we summarize it," requires about 50 gigabytes, "and we think by the end of the year, we will probably be in the range of about 10G," Flatley said.
In addition, the company formed a "genome informatics alliance" last year, he said, to study ways of analyzing vast amounts of sequence data. "We called together all the leaders in the IT field, the Amazons, the Googles and Microsofts, as well as the genome centers [and] customers from the high-energy physics world, to really study this problem and get prepared for the onslaught of sequencing data that will be coming here in the next few years," Flatley said. The firm also has a "very active" cloud-computing initiative, he added, and believes that its customers "will increasingly put these datasets into the cloud and take advantage of fully scalable computing."
Flatley also commented in his presentation on a "new phase" for genome-wide association studies that he believes will continue to use genotyping arrays with novel content from the 1000 Genomes Project instead of sequencing.
Sequencing, he said, "is not an applicable technology" for such projects because it is still too expensive. "Arrays are 100 times cheaper, 100 times faster, and materially more accurate than sequencing and will be over the next couple of years," he said.
Pre-announcing preliminary earnings for the fourth quarter of 2009, Flatley said that the company had approximately $176 million in total revenue. This translates to 2009 revenues of about $662 million, a 15-percent growth over 2008. In 2010, Illumina expects revenues to grow another 20 percent.