Illumina, Inc. (ILMN) Q2 2012 Earnings Report, Transcript and Summary

Kevin Williams Marc Stapley - Chief Financial Officer, Senior Vice President and Principal Accounting Officer Jay T. Flatley - Chief Executive Officer, President and Director Christian O. Henry - Senior Vice President and General Manager of Genomic Solutions

Tycho W. Peterson - JP Morgan Chase & Co, Research Division Doug Schenkel - Cowen and Company, LLC, Research Division Amanda Murphy - William Blair & Company L.L.C., Research Division Derik De Bruin - BofA Merrill Lynch, Research Division Ross Muken - ISI Group Inc., Research Division William R. Quirk - Piper Jaffray Companies, Research Division Daniel Brennan - Morgan Stanley, Research Division Nandita Koshal - Barclays Capital, Research Division Isaac Ro - Goldman Sachs Group Inc., Research Division Daniel L. Leonard - Leerink Swann LLC, Research Division Daniel Arias - UBS Investment Bank, Research Division Peter Lawson - Mizuho Securities USA Inc., Research Division Adam Darity Zarak Khurshid - Wedbush Securities Inc., Research Division Paul R. Knight - Credit Agricole Securities (USA) Inc., Research Division

Good day, ladies and gentlemen, and welcome to the Second Quarter 2012 Illumina Incorporated Earnings Conference Call. My name is Melanie, and I'll be your coordinator today. [Operator Instructions] As a reminder, today's call will be recorded. I would now like to turn the call over to Mr. Kevin Williams. Please proceed.

Thank you. Good afternoon, everyone, and welcome to our earnings call for the second quarter of 2012. During the call, we will review our financial results released today after the close of the market and offer commentary on our commercial activity. After which, we will host a question-and-answer session. If you have not had a chance to review the earnings release, it can be found in the Investor Relations section of our website at illumina.com. Participating for Illumina today will be Jay Flatley, President and Chief Executive Officer; Marc Stapley, Senior Vice President and Chief Financial Officer; and Christian Henry, Senior Vice President and GM of our Genomic Solutions business. This call is being recorded, and the audio portion will be archived in the investor section of our website. It is our intent that all forward-looking statements regarding our expected financial results and commercial activity made during today's call will be protected under the Private Securities Litigation Reform Act of 1995. Forward-looking statements are subject to risks and uncertainties. Actual events or results may differ materially from those projected or discussed. All forward-looking statements are based upon current information available, and Illumina assumes no obligation to update these statements. To better understand the risks and uncertainties that could cause actual results to differ, we refer you to the documents that Illumina files with the Securities and Exchange Commission, including Illumina's most recent Forms 10-Q and 10-K. Before I turn the call over to Marc, I want to let you know that we will participate in the Morgan Stanley Global Health Care Conference in New York the week of September 10 and the UBS Global Life Sciences Conference also in New York the week of September 17. For those of you unable to attend these conferences, we encourage you to listen to the webcast presentations, which will be available through the Investor Relations section of our website. With that, I will now turn the call over to Marc.

Thanks, Kevin. Good afternoon, everyone, and thank you for joining us today. During this section of today's call, I will review our second quarter financial results and our guidance for the remainder of the year. I will then turn the call over to Jay to provide an update on our commercial progress and the state of our business and market. We have continued our positive revenue and earnings momentum with our third consecutive quarter of sequential growth in both revenue and earnings per share. In Q2, we had record operating margins and record orders of the both the MiSeq instrument and MiSeq sequencing consumables. Second quarter 2012 revenue increased 3% sequentially to $281 million due to another successful quarter of MiSeq shipments, higher service revenue and continued growth in sequencing consumables. Revenue for the quarter decreased 2% compared to Q2 2011 as last year's quarter included a greater number of HiSeq sales. Instrument revenue for the second quarter was $72 million compared to Q2 2011. Instrument revenue was down 32%, again primarily due to decrease in sequencing and microarray instrumentation. Consumable revenue for the quarter was $184 million, up 6% sequentially and up 16% compared to the second quarter of 2011. Consumable revenue now represents 65% of our total revenue compared to 55% this time last year. We are very pleased with this trend, which results from a larger install base of both HiSeq and MiSeq and a sequentially improved average annualized quarter on HiSeqs. In particular, record shipments of our [indiscernible]product including TruSeq and Nextera contributed to the sequential growth consumable revenue. Services and other revenue, which includes genotyping and sequencing services, as well as instrument maintenance contract, was $22 million for the quarter compared to $18 million from Q2 of last year. This growth is also encouraging. In our discussion of gross margin and operating expenses, I will highlight our adjusted non-GAAP results, which exclude non-cash stock competition expense, restructuring charges and in-process R&D charge and other non-cash items. I encourage you to review the GAAP reconciliation of non-GAAP measures included in today's earnings release. Our adjusted gross margin for the second quarter was 70.9%. This compares to 69% in both last quarter and the second quarter of 2011. The sequential improvement was attributable primarily to a favorable mix of consumable versus instrument revenue and improved absorption. The year-over-year gross margin increase was also driven by a favorable product mix. We continue to expect gross margins of approximately 70% for the full year as improvements in consumable mix are somewhat offset by expected MiSeq care [indiscernible] reduction as our trade-in program continues to gain traction and by increasing FastTrack Services revenue. Adjusted research and development expenses for the quarter was $41 million or 14.7% of revenue compared to 14.9% of revenue in the first quarter and 14.1% of revenue in the second quarter of 2011. Adjusted SG&A expenses were $55 million or 19.4% of revenue compared to 18.9% of revenue in the first quarter and 19.1% of revenue in the second quarter of 2011. Adjusted operating margins were a record 36.7% of revenue in the quarter compared to 35.8% of revenue in the second quarter of 2011 as we gain leverage from the improved gross margin performance. Our non-GAAP tax rate for the quarter was 33.8% compared to 34.7% in the second quarter of last year. This quarter's tax rate was higher than the expected annualized rate as a result of the delay in the passage of the U.S. R&D tax credit. We anticipate the non-GAAP tax rate will continue to be approximately 34% until the R&D tax credit is passed and retroactively applied at some point this year. Non-GAAP net income was $53 million for the quarter, and non-GAAP earnings per share was $0.40. This compares to non-GAAP net income and EPS of $52 million and $0.38, respectively, in the second quarter of 2011. We reported GAAP net income of $23 million or $0.18 per diluted share in the second quarter compared to $31 million or $0.22 per diluted share in the prior-year period. This includes an impairment charge this quarter of $21.4 million related to a technology asset acquired in 2010 and $6.7 million of ongoing unsolicited tender offer expenses. During the second quarter, we generated cash flow from operations of $96 million. We used approximately $21 million for cash capital expenditures resulting in $75 million of free cash flow. This compares to $55 million of free cash flow in the second quarter of last year and $52 million in Q1 of this year. DSO decreased to 61 days compared to 67 days last quarter and 61 days last year. And we ended the quarter with approximately $1.3 billion in cash and short-term investments. During the quarter, we repurchased approximately 800,000 shares for $32 million. With these purchases, we have slightly over $200 million remaining in our authorized share buyback program. We are pleased with our financial performance for the first half of 2012. While some uncertainty exists with respect to academic and research funding in the second half of the year, our outlook is generally as we anticipated. Accordingly, we are reaffirming our 2012 revenue guidance and increasing our non-GAAP earnings per fully diluted share expectation from $1.40 to $1.50 to now be between $1.50 and $1.60. At this point, I will turn the call over to Jay for some remarks on our commercial activity during the quarter before we begin the Q&A session.

Thanks, Marc. We are very pleased with our second quarter results. The positive trends that began after Q3 2011 have continued through first half of 2012. We had sequential revenue and EPS growth with record operating margins, and we believe we positioned the company well in light of the macro uncertainty expected in the second half the year. While there's no clarity on sequestration and the NIH budget for 2013 remains uncertain, we continue to believe that the worst case 8% funding reduction is an unlikely scenario. The Obama administration has proposed a flat NIH budget for 2013, while last month, the Senate Appropriations Committee approved $100 million increase to the NIH budget for next year. More recently, the House Committee on Appropriations released their fiscal year 2013 Labor, Health and Human Services Funding Bill, which proposed roughly flat NIH funding year-over-year. While these are all positive indicators, we don't expect any budget clarity until a few months after the election. We also believe that our U.S. academic customers are better prepared than last year for budget challenges and are planning accordingly. In light of that, it's our view that the overall funding dynamics this year are considerably more favorable than last year, and we've considered a range of potential budget outcomes in formulating our annual guidance. We remind investors that in Q1 2012, our core sequencing consumable revenue was for the first time ever greater than instrument revenue, and this continued in Q2. A greater mix of consumable to instrument revenue provides more inherent visibility on our business. To improve this visibility even further, as we discussed in our Q1 earnings call, we proactively worked with some of our customers to project their consumable demands for the rest of the calendar year. By agreeing upfront to take periodic noncancelable, consumable shipments, these customers avoided our Q2 price increase. Tremendous growth in our sample prep products also improves this visibility given their highly recurring use and our customers' workflow. We're also seeing a continuing evolution of our customer mix toward more commercial, clinical and applied accounts reducing our exposure to academic funding. To further enhance our portfolio and product diversity, we announced a number of new offerings during the quarter, including Nextera XT DNA Sample Prep kits, the BaseSpace Apps store, the iSAAC genome alignment tool and a RapidTrack Whole Genome Sequencing Service. Additionally, we have numerous products that will begin shipment in the near term including the HiSeq 2500, the associated HiSeq 2000 upgrades, enhanced MiSeq, a 2 by 250 base pair of MiSeq sequencing kit and lower throughput kit, NuPCR consumables and new sample prep products like Nextera Exome and Custom Arrangement kit. While our overall view in the second half of the year remains cautious, with improved forecast visibility, new products and services launches, the change in customer mix, improved customer planning and fewer funding dynamics at play, we believe that we're in considerably more favorable position this year relative to last year. Let me turn now to some specific results of the quarter. Shipments during the quarter to the Americas were healthy with U.S. shipments up 18%, Canadian shipments up 15% and Brazil shipments up over 100% year-over-year. Since opening an office in São Paulo in Q1 on 2011, Brazil has moved up to be a top 12 country as measured by x U.S. shipments. With regard to Europe, we've not seen a material change, and our business there remains stable overall. Shipments to Spain and Italy actually increased sequentially during the quarter. Shipments to Asia, while down sequentially, grew year-over-year and remain healthy. This sequential decrease was expected as Q1 is the end of the Japanese fiscal year. Notable strength was seen in China where shipments grew 50% year-over-year. In the array business, total microarray revenue was roughly flat sequentially. This quarter marked the second-highest number of Infinium samples ever shipped. We received orders for approximately 250,000 more Exome samples bringing total samples order to date to approximately 1.6 million, by far the highest number of samples for any content collection in the company's history. New unit orders for iScan and HiScan microarray instrumentation were the highest since Q2 2011. In addition, Q2 saw market sequential orders strength in our Omni express, as well as our Ag arrays, including Bovine LD, PorcineSNP60 and MaizeSNP50. Our CytoSNP array shipments had their second best quarter ever and remain on track for submission of our iScan with CytoSNP for FDA 510(k) approval later this year. Yesterday, we launched our real-time PCR reagent portfolio, including a novel probe-based chemistry for gene expression analysis that we call NuPCR. In addition to the NuPCR reagent, we announced qPCR DNA-binding assays for gene expression and qPCR High Resolution Melt assays for genotyping studies. These provide researchers with a trio of high-quality, budget-friendly options in the qPCR reagent market. This reagent portfolio will be sold through our worldwide distribution network. Availability of NuPCR in the United States will be announced in the near future. NuPCR is a novel probe-based technology that doesn't rely on exonuclease activity to generate signal. Instead, the catalytic activities inherent in the 3-part oligo probe resulting in higher sensitivity experiments and consistently detect targets earlier in the cycle count and competing real-time PCR probes and it reduced cost. This collection of new assays is compatible with any real-time PCR platform offering researchers a wider range of choices. Additionally, all the assays are custom-designed using DesignStudio, a personalized web-based design tool available to Illumina customers. This unique easy-to-use software delivers bioinformatically optimized assays, ensuring high PCR efficiencies and assays that work the first time without tedious trial and error optimization, while also enabling simplified multiplexing. Turning now to our sequencing business, total sequencing revenue in Q2 grew mid-single digits sequentially attributable primarily to growing sequencing consumables, as well as growing service revenue for maintenance contracts on our larger install base. Total sequencing consumable revenue grew over 35% compared to Q2 of last year. We had record HiSeq and MiSeq consumable shipments during the quarter, and the average flow-through on HiSeq again increased sequentially. TruSeq and Nextera sample prep revenue year to date has grown 57% year-over-year demonstrating our focus on back-integrating into sample prep. We expect continued growth here as researchers convert from homebrew methods. We continue to gain competitive market share. We launched new sample prep product and our install base grows. Average annualized consumable utilization of HiSeq increased to $338,000 in Q2 from $299,000 in Q1, and we have several reasons to feel confident that this level of utilization can continue as the effect of the consumable price increase announced in April should be more greatly reflected in future quarters. Additionally, in Q4, we will begin to see the impact of the rapid run mode of the HiSeq 2500, which yields greater revenue per genome, and we expect to see sample prep revenue continue to grow. During the quarter, we introduced the Nextera XT DNA sample prep kit, the easiest way for researchers to prepare and sequence low input small genomes, PCR amplicons and plasmids. Compared with MiSeq, Nextera XT provides the fastest timed result of any next-generation sequencing technology, enabling researchers to go from genomic DNA to analyze data in less than 8 hours. Nextera XT will also be available for HiSeq allowing up to 1,536 samples per HiSeq 2000 run and up to 384 samples per HiSeq 2500 in rapid run mode. Last week, we introduced Nextera Exome and custom enrichment kit, leveraging the speed of Nextera technology and supporting the industry's lowest DNA sample input requirements, the new kits enable researchers to quickly and economically perform a wide range of studies from small focused gene panels to full human exomes. It also enable researchers to study small samples while retaining sufficient material for future analysis and offer the ability to process large numbers of samples without investing in automation systems. Sequencing instrumentation revenue was down sequentially primarily as a result of reduced MiSeq shipments as the backlog of MiSeq orders from 2011 was mostly cleared in Q1. MiSeq ASPs were also down slightly during the quarter, largely a function of our continuing successful trade-in program. For the first time since the second half of 2010, HiSeq 2000 orders and shipments increased sequentially. ASPs of HiSeq 2000 shipped in Q2 also rose sequentially. Through Q2, approximately 75% of new HiSeq 2000 orders have included an upgrade package. And as we begin to ship the new HiSeq 2500s in early Q4, the promotional upgrade price for the install base of HiSeqs will end, rising in price from $50,000 to $125,000. With the shipment of the 2500, we will provide the ability for all HiSeq to stream data to BaseSpace removing the requirement for local IT and bioinformatic capability. We continue to be very pleased with the market penetration of MiSeq. Customer feedback continues to be very positive, and the independent researchers have demonstrated, in peer-reviewed publications, the superiority of MiSeq over other desktop sequencers with regard to data quality, ease-of-use and end-to-end workflow. We're also pleased with the success of our trade-in program with such orders doubling sequentially from Q1. During the quarter, customers placed record orders for both MiSeq instruments and consumables. And MiSeq consumable shipments grew almost 60% sequentially. The average annualized flow-through in Q2 on MiSeq was approximately $55,000. The upgraded version of MiSeq will begin shipping from our factories within the next 2 weeks and the field upgrade program is about to begin. Initial test upgrades have gone extremely well. One of our early access customers has seen runs with throughput of approximately 10g with the data quality our customers have come to expect. With the new 2 by 250 base pair kits available on MiSeq, we expect aggressively market against longer read desktop machines and convert those users to MiSeq. More than half the orders for MiSeq continue to come from non-academic customers. Finally, we remain on track to submit both the MiSeq platform and our own proprietary assay to the FDA for 510(k) approval before year end. In our FastTrack Services business, we had record orders in Q2 resulting from booking approximately 3500 genomes as well as a large microarray service contract. We expect our success in genome bookings to continue in Q3. During the quarter, we launched the RapidTrack Full Genome Sequencing Service, a new offering that will deliver a whole human genome in less than 2 weeks. This premium service offers the fastest sample to data turnaround time to any whole genome sequencing service available today. Currently, we expect this service to be interesting to those research groups that needs samples analyzed quickly, to meet project deadlines or to submit grants and proof of principle studies. We've received several orders testing our RUO process with an eye towards qualifying this service under CLIA before the end of the year. During the quarter, we introduced BaseSpace Apps, the dedicated application store for BaseSpace, our genomics cloud computing platform. Informatics solutions available through BaseSpace Apps will allow customers to leverage a growing community of academic, commercial and open-source tool providers for building applications around the Illumina data to dramatically simplify and accelerate genomic data analysis. Initially, 14 early-access vendors are developing apps for BaseSpace with more than 20 other vendors working to deploy in BaseSpace tune as a developer portal and API are released publicly later this quarter. This morning, we announced our BaseSpace pricing for both data storage and downstream analysis. This is a first in the NGS market where customers will have the opportunity to receive free data storage and analysis on all Illumina data and have the ability to augment their account with flexible storage and app purchases. Our demonstrated leadership in informatics enables us to provide effective solutions for those lacking informatics infrastructure or expertise and allows us to open a new market opportunity. This informatics ecosystem will greatly simplify data analysis during and after our researches data has been generated. The data is always accessible and secure no matter where the research is located, and the App Store offers a wide range of bioinformatic tools and application. By greatly simplifying the management analysis of data, researchers can focus on the main task: understanding the biology. We're also excited to announce a growing adoption of BaseSpace along our MiSeq users with a vast majority of MiSeq systems now connected to BaseSpace and roughly half of those actively uploading data, demonstrating that our architecture's resonating in the market. This fall, Illumina will hold a company-sponsored symposium that we're calling Understand Your Genome to explore the best practices for deploying next-generation sequencing in a clinical setting. Illumina plans to sequence and analyze the genomes of approximately 60 participants as a multi-step process to engage experts in the field around whole genome sequencing. The first day will feature presentations on clinical, laboratory, ethical, legal and social issues around WGS. The second day, participants will receive their genome data on an iPad and learn under the guidance of the medical geneticist how to analyze their results using their MyGenome app. Through this program and related efforts, we're laying the groundwork for routine clinical use of whole genome sequencing, clearly the wave of the future. In conclusion, we're very pleased with our execution in the first half of 2012, and we're excited about our product lineup for the remainder of the year. We sequential revenue in EPS growth and record operating margins resulting from our focus on strong operational execution and continuous innovation to maintain our technological lead. And we continue to believe that the overall sequencing market has enormous potential and that we have the technology, people and infrastructure to maintain our market leadership and capitalize on this unique opportunity. Thank you for your time. We'll now open the lines for questions.

[Operator Instructions] And our first question comes from the line of Tycho Peterson with JPMorgan. Tycho W. Peterson - JP Morgan Chase & Co, Research Division: Maybe the first question on the PCR announcement last night and I noticed in there, your talk about disruptive pricing. So I'm wondering if you're able to talk at all about your pricing strategy there. And should we expect a U.S. distribution partner for that business in the back half of the year? And lastly, can you talk a little bit about, do you need to expand the instrument lineup for your own business beyond Eco?

There's 3 different assays Eco that we announced. The principal one that we think will have the biggest market impact is the probe-based assays that we call NuPCR. And those we are pricing at disruptive level. So there'll be priced at about 50% of what the traditional PCR assays would be priced at in the overall market. It's a little bit complicated because there's a couple of different components. There's a master mix part that gets amortized over how many samples you run, but the probe-based part will be priced very aggressively. So we think that it has a great opportunity. And the fact that it runs across any install PCR instrument, we think, will be a big help. In terms of distribution, we are working on our U.S. distribution strategy. We have some interesting progress there, but nothing yet to announce today. So hopefully, in the back half of the year, we certainly plan to have some announcements about the U.S. distribution strategy. With respect to instrumentation, we're continuing to focus on this as a business unit, and we certainly are looking at ways to continue to involve the existing instrument as we do with most of our platforms and also clearly exploring what other new instruments we might put into the PCR space. Tycho W. Peterson - JP Morgan Chase & Co, Research Division: And then turning to MiSeq, are you able to talk about system reliability in the field? There were some valid issues which obviously I think you worked through. But just are you able to talk about more broadly the downtime on MiSeq today? And I think Marc had made some comments earlier about discounts and instruments continuing. I'm wondering if there's anything more there in terms of pricing slippage? Or was that just reflective of the trade-in improvement?

Yes. We've been very pleased with the overall reliability, Tyco. But we have had one valve problem, the selector valve in the MiSeq. We've identified the root cause of that problem, had to do with particulates in the valve that was coming from our vendor. There's a couple of things we've done there. We -- the vendor now goes through an extensive wash protocol before they ship the valves to us. And these new valves are incorporated in all new shipments and have been for about a month, I think, now. We also introduced a wash protocol, so that anybody who has an instrument already, whenever the instrument runs or when it completes a run, we wash the valve out so that any particulates that were in the valve don't remain, because they essentially create nucleation sites for particulates, and that's what was freezing up the valves. And then as we go into these MiSeq upgrade in the field to the enhanced version of the MiSeq, we're going to change out any of the old valves and make sure all the instruments have brand new valve. So we think that field issue is totally taken care of. With respect to pricing, what we mentioned in the script really didn't have to do with discounting. The reduction ASP had to do with the trade-in program. I guess you could call that a discount in some way, but it's specifically around the trade-in program. So we have trade-ins for competitive instruments. And in some cases, we've taken some trade-ins for GAs as well. And in those cases, we've offered a credit off of the list price of MiSeq active customers. But other than that trade-in program, if you take that out of the numbers, the pricing for MiSeq is holding quite well. Tycho W. Peterson - JP Morgan Chase & Co, Research Division: Okay. Last one for me. Any color you can provide on orders or book-to-bill? We've worked it in the question as you can imagine from investors, so just wondering if you're willing to comment on either of those?

Yes. I mean, we don't want to get into a regular pattern of reporting book-to-bill. That hasn't been our typical pattern. We did talk about it the last 2 quarters because we were in the middle of the RUO situation, and there were question about whether we were pulling in revenue, and we felt that we had to disclose book-to-bill as a result. What I can say is that we are very pleased with our incoming order rate, that we exceeded our forecast, and we're very happy with the business. And as I indicated in the script, we are not seeing, at least as yet, any impact of funding concerns, but we remain cautious about the back half.

Our next question comes from the line of Doug Schenkel with Cowen & Company.

So the increase in HiSeq annualized utilization is $338,000, I believe you said. It's quite impressive. Does that revenue -- does that number include revenue associated with Nextera, and more broadly, your efforts to capture front-end revenue on sequencers? And I guess what I'm trying to get at is how much of this increase is that versus price hikes versus actual year-over-year increases in sequencing utilization?

It does include the sample reagents, Doug, so that is included in the numbers. If you turn the clock back a year, our sample prep revenue was a much smaller fraction, so it was less significant last year. So it is a component of that growth, which we think is fine. We'll take it however we can get it. But it certainly also, I think, is an indication of the overall increased utilization rate and a little bit of price increase. But most of the price impact I think you're going to start seeing next quarter. There's probably some in Q2 as well. But you'll probably see the full impact of the price increase in Q3.

But if you didn't have that initiative to get more of the front-end spend, would the utilization still be picking up quarter by quarter?

Yes, it would, but not as dramatically.

Okay. You kept operating spend about flat in dollar terms relative to Q1. Your guidance doesn't seem to incorporate a material increase in operating spend in dollar terms for the second half relative to the first half. You have a series of new products launching as you mentioned in your prepared remarks. I'd expect you're going to be rolling out new products over the balance of the year. You have aspirations in diagnostics amongst other areas. How long can you keep the spend at these levels, especially keeping in mind the streets are looking for 8% growth in '12 and '13, and clearly, folks would like more?

Yes, I mean, we definitely are planning some expense increases in the second half of the year, so I think the assumption that expenses would be flat in dollar terms isn't right. I mean, obviously, I think we'll try to keep them in the range as a percentage of revenue that we've been running in. But in dollar terms, we do expect to increase those expenses. We were really pleased with our gross margin performance this quarter, so 70.9% is not as high as we've hit. Historically, we've hit a little bit higher than that, but it's pretty close to our historical all-time high. And we're feeling pretty positive about our ability to continue to push on the margin line. So I think from that perspective, we're feeling pretty comfortable about being able to increase expenses next year and hit the revised EPS that we put out there.

And, Jay, I think you said you're going to have the MiSeq and the proprietary assay at the FDA by year-end. Could you provide any more detail on what that assay is going to be?

We can't as yet, Doug, so we haven't disclosed what it's going to be. But it's -- I would describe it as an assay that will be unique to us in there, but doesn't have proprietary content. So this is not content that we have unique intellectual property around. Let's be clear about that.

Okay. And maybe if I can sneak in one very last one. How was your outlook for HiSeq placements evolved over the last couple of quarters given that it sounds like you had actually a pretty good HiSeq quarter for placement?

Yes, we did. We've been really pleased with the impact the 2500 has had on the HiSeq trends. As we mentioned, about 75% of the incoming orders include the upgrade, so this idea that there's not money for capital is really not playing out here because it costs a significant more for customers to get that upgrade when they place the order. So I think that, that has helped our outlook on the overall sort of trend line for HiSeqs. We're actually seeing surprisingly good performance in the 1000 and 1500 as well. So we're probably doing a little better there than we might have forecast a year ago. So we're not ready to talk about extensive growth in the instrument number yet on HiSeq, but it grew sequentially, and we're really pleased with that, and we're obviously going to keep trying to move it up.

Our next question comes from the line of Amanda Murphy with William Blair. Amanda Murphy - William Blair & Company L.L.C., Research Division: I actually had a follow-up question on the guidance. Just looking at the revenue guidance that you've reaffirmed and just given the visibility that you have relative to last year in terms of the consumable usage. Can you help us frame out sort of how to think about the range and growth in the back half? Is that primarily funding driven? Or is there anything else that's key there?

Yes, Amanda, it's Marc Stapley here. We've got quite a broad range still for the second half given the uncertainty that Jay referred to in his prepared remarks. So we're thinking about -- there are a number of scenarios. Obviously, the bottom end of that range will represent a fairly flat revenue scenario in the second half and the top end of the range is obviously some fairly substantial growth. So we've modeled out those various scenarios. It clearly is funding driven. As we say, we haven't seen any effect on our international order rate at this point, but we still remain cautious there. And on the U.S. funding, we've got no more certainty than we've had with respect to that in the first half. So that's really how we're thinking about it. And our outlooks have been spending fairly well so far. Amanda Murphy - William Blair & Company L.L.C., Research Division: Got it. And then just thinking about HiSeq consumable usage, just putting the price increase aside for a second. Can you just help us think about how the usage is coming across the major user bucket, so think about PGI versus the core labs versus genome centers? Just trying to get a sense of how that number can trend over the next few years and sort of how much is this coming from the smaller labs' increasing usage? And if you think about BaseSpace and some sample improvements or how important our data that 338 number going up over time.

I'd say in our largest customers, so BGI and the genome centers, that they still have some excess capacity. They drop back considerably during the 3Q challenges of last year when the NHGRI funding was up for grabs. They've come back from that level for sure but they still, we think, have some excess capacity. One of the things that is growing the number we think is that most of the incremental HiSeq orders we're getting at this point are from customers who already have a HiSeq. And what that implies is that they maxed capacity on the 1 they have or the 2 or the 3 that they have, and they're some incremental unit. And when that happens, typically, the overall utilization goes up because these people are in production, those customers are in production. So I think that's one very positive trend. Obviously, the pricing helps. The increased market share in sample prep helps. All those factors I think are potentially things you could argue might be able to nudge that number up a little bit. But because of the funding issues overall and the experience of last year, we're certainly not projecting it to go up because there could be some funding challenges in the back half. So those are the factors to consider when you try to model out where it might go. Amanda Murphy - William Blair & Company L.L.C., Research Division: Okay, and then just switching to the MiSeq and the trade-in program that you talked about. Just curious if you can provide a perspective on how many, I don't know if it's boxes or sales sort of are -- I guess boxes more, were trade-in and how long do you anticipate running that program for?

It's open-ended, so we don't have a defined end date to it. We may put one on there at some point in time, but we don't have one now. I can say the quarter was in the range of 20-ish trade-ins overall across a mix of competitive instruments and a couple of our older GAs as well where we gave some trade-in credit. We -- the pipeline, as we look forward for potential additional trade-ins, is pretty rich. So we'll see what happens in Q3, but the potential is for the number to be at least higher, maybe higher.

I also think with the performance of the MiSeq and the upgrades, as Jay pointed out, the early access customers have been getting higher throughput than maybe we advertised, that might also create new demand -- incremental demand for upgrades.

And one other thing we might mention that we didn't include in the script is that all MiSeq that ship out of the factory in Q3 and later will be the enhanced MiSeq. So we finish shipping all of the prior generation of MiSeq at the end of Q2.

Our next question comes from the line of Derik De Bruin from Bank of America.

So all -- the array business looks like it had a little bit of a resurgence this quarter. Could you just talk about it? Is this just a tick up in the studies? Is this share gains? Can you just put a little bit more color on the array market? And also kind of put it in scenes of what you see in terms of converting expression arrays to RNA-Seq and how it kind of fits with the whole exome kind of analysis.

Yes. I mean, there's a lot of factors at work here. We would characterize the array business overall as roughly flat. We had a positive trend on the instrument side during the quarter in the array business. In terms of the sub pockets of that business, I think whole genome association type chips are trending downward, but that's made up for by the positive trend lines in the exome chip, where we continue to have tremendous success. And we don't put that in the whole genome category. We put that in our focus category, and we've had really positive trend lines from some of the Ag chips as well, and the custom business continues to do well. So I think you can think of those factors roughly as netting out in Q2.

Okay. And talking about your services business and 3500 genomes, it's like what's the pricing looking like in that? And it's like, how do we think about the order book for that? Are you seeing more outsourcing? Are you picking up more contracts? Just can you talk about this? Particularly as you kind of launch new products, you're going to need faster turnaround.

Well, we're seeing a lot of those genomes come in that are cancer genomes. And actually a fairly high percentage are from cancer sites. And many of these are nontraditional sequencing centers, so they're places that you wouldn't typically expect to have installations of 5 or 10 HiSeqs. The other kind of deal we see are ones that want relatively fast turnaround and don't have an existing infrastructure. So they don't want to buy instruments and take 6 months to set up a sequencing factory. They want to get all their samples in and turned around in 3 or 4 months, and I think that's driving it in part. In terms of pricing, it's probably down a little bit in overall ASP from what we saw over the past couple of quarters, but I think it's beginning to flatten out some.

Okay. And just want to clarify on the -- I just want a little bit more color on the MiSeq business. So you have seen the 2 by 250 sample kit coming in for MiSeq. I guess is that for the version 2 system, the version 1 systems? I'm just -- and sort of what's the turnaround time inclusive of sample prep for running that?

Yes. It runs on the version 2 system. The overall runtime of the sequencer, Christian, you know that number is?

It's over 24 hours, I think. Yes.

So it's not 20?

It's not 20. We don't even know exactly but it's --

40? We're getting a signal that it's 40 hours.

40.

Okay.

We get the 2500 in 27 hours and the upgrade of MiSeq in 48 hours.

Yes. So we've increased the cycle times on the MiSeq too, so it goes -- per cycle it goes faster, but there's a lot more cycles here so...

Right, 250, right. So -- and I guess also on the MiSeq, I mean, you said that some of the instrumentation slowed down -- the instrumentation revenue decreased sequentially due to the backlog coming out there. Can you quantify that?

Well, I can qualify it. I'm not sure I can quantify it for you. But so what happened, obviously, we had -- we're taking orders for a couple of quarters before we began shipping. And as we're ramping the manufacturing lines, we began to whittle down that backlog. But in the first quarter, we started shipping it. The incoming order rate was about equal to what we shipped, so we actually didn't make a lot of progress on getting the backlog reduced. And over the subsequent 2 quarters, we did that. And in Q1, we came out of that quarter with sort of a normalized backlog and had just a normal Q2 in terms of what we'd expect, where the shipments about match the incoming order rate. But we actually had a great order quarter from MiSeqs as well, so it's continuing to grow. And we're now in a place where we can deliver MiSeqs in a kind of time frame that our customers expect. And as we said during the time with the HiSeqs, customers don't want to wait 2.5 months to get these instruments, so we want to deliver them in 4 weeks or so, sometimes a little less if we can. Yes, I think the key metric is increasing MiSeq orders each quarter. That's the key metric.

Our next question comes from the line of Ross Muken with ISI Group.

I wanted to wanted to walk through, Jay, your initial comments kind of on the sequester and sort of the state of funding in the U.S. I think the growing view at least on the Wall Street side of things, I think on the political side of things is there's a high probability at least this sequester will come into some place Gen 1 starting next year. The question is more sort of the length of period that which it's actually in place and then readjust it to kind of a new normal budgetary environment. Assuming that, that is a scenario, how do you think customers start to respond from a purchasing perspective in the back half of the year? And what's the sort of differential in terms of how you would expect a large versus a small customer to sort of swallow the potential temporary change in the environment?

Well, I mean, what we think is going to happen is that we're going to come out of this year, fiscal year, so the beginning of October with a continuing resolution into Q4 because any chance of us having any real budget is close to 0, I think. And so if there's going to be some continuing resolution that's probably going to have a negative number in it. And what that is anybody's guess, but that there will probably be some controlled spending in that continuing resolution through Q4. And probably nothing's going to happen until after the election. And so that'll continue through the time frame. And then depending upon who gets selected, we'll have some budget resolution we think in the first month or 2 of Q1, of calendar Q1 2013. What happens between January 1 and when we actually get a budget? I don't know. I guess there's a chance it go to the 8% sequestration. But I'd be surprised. I would guess that whatever the continuing resolution is, we're going to continue all the way through to the point where we have a real budget.

Okay. Maybe it's a nuance, but the difference in the actual core budget and the sequesters are sort of 2 separate entities, right? So to the degree, the sequester occurs that sort of on top of whatever happens with the budget. And so that is an automatic cut until the Congress does something to relieve it. And so I guess what I'm trying to say, let's just assume that, that will be the magnitude. I mean, how do you think customers start to react? Do you think your customer would preference certain types of research? Because it only affects new grant cycles, correct? It wouldn't obviously affect the legacy grants. I'm just trying to get a sense from a risk-profile perspective, like what do you think the order of operations is for a lab manager of one of your typical customers in terms of having to adjust to an environment that could be strained even for a small period of time.

Yes. Well, I think they would get a little more conservative during that window because they'd be worried -- even though they're working off existing grants, they'd be worried about how much future money they're going to get and would be thinking about potentially stretching it out. We think that's already happening to some extent. And I think our customers are thinking about that issue today, not all of them, but quite a number of them are and are sort of metering out the dollars in anticipation of a little bit of disruption in the back half.

Okay. And just sort of the backlog commentary, I know you don't want to give too much. But just so understand at least from like a MiSeq modeling perspective. So the way to think about is sort of like what happened with HiSeq. We've now gone through most of the backlog. Were getting now to kind of the new normal level of placements. And then Christian's comment, the key to sort of lodge will be the sequential improvement, order rates, which will now sort if drive the instrument growth sequentially. Is that sort of the right way to think about it?

Well, yes, that certainly. But also the increasing reagents because of the growth of the install base will be accurate.

No, I'm just thinking from a capital equipment perspective. So we sort of gotten the new normal and now we've got to grow sequentially vis-à-vis orders versus there being a significant backlog in effect.

That's right. And the one thing to think about there, we didn't include in any of our prepared remarks, but if you look at the mix of MiSeq orders to date, the vast majority of these are single unit orders. I mean, very, very high percentage. And what that means is that we have lots of customers who are developing assays, developing clinical tests, setting up research protocols on these units. And what we expect to see are the beginnings of repeat orders from those customers as they begin to fill up MiSeq number 1 and move on to MiSeq 2 or 3. And so the need to continue to find new customers is not quite as high to continue to grow that number because of that beginnings of the repeat business which we think we still have.

Our next question comes from the line of Bill Quirk with Piper Jaffray.

This question, Jay, just to, I guess, the risk of beating a dead horse here on spending. I just want to, I guess, elaborate a little bit on the comment regarding not seeing a lot of pressure here. Can you talk at all whether or not you're seeing any re-prioritization? In other words, reemphasizing consumables versus seeing new instrumentation and all?

No, we haven't. I ping our sales VP every week now on this looking for any early warning signals of changes in customer behavior, and were just not seeing it yet. And maybe it's too early because I think they feel probably pretty good with respect to what's going to happen through the end of Q3 -- or the end of Q4, I mean, -- Q4 fiscal, so our Q3. And there might be a little bit more effect in Q4. But so far, we're just not seeing any changes in customer behavior. We're certainly on the lookout for it.

Of course, the way we look at our sales model, it's based on the current funding opportunities. And when you look at a product like MiSeq, the funnel is very strong.

And then just thinking a little bit about sort of MiSeqs that are going outside of academia. Do we have enough data points at this stage to talk about what the utilization looks like within this customer base as compared to, say, each of your core academic group?

I don't think it would be very meaningful yet. And the reason is that many of those are going to clinical accounts where they're in the stage where they're doing the clinical research work, they're not actually running them in a production environment yet. And so I don't think that -- I don't have those numbers at my fingertips, but even if I did, I don't think they'd be necessarily indicative of what that segment will get to, say, a year from now.

Understood. And then just last quick one for me. It's just regarding the comment around the customers that you've signed on to receiving periodic orders on the consumable side, so as to avoid the price increase, can you talk at all to what extent that'll influence the back half of the year? To what extent that you've you been able to say -- communicate that as a percentage of what your overall consumables expectations will be?

Well, I'd say of the total consumables for the next 6 months, it's not a significant fraction of that total. But it's something that helps us forecast, and it also tends to be something that occurred obviously more with our larger customers. So it gives us visibility on the behavior of those bigger customers. And that's quite useful, and they're the ones that perhaps were the most interested in avoiding this price increase.

Our next question comes from the line of Daniel Brennan with Morgan Stanley.

Just a question starting with HiSeq. I believe in the past, you've talked about maybe 25% to 30% of the HiSeq user base expected to be upgrading to the 2500. Is that still the case? Since it looks like certainly in the new orders, you're seeing a very high percentage of those take the option to upgrade. I just want to get a sense of how we should think about the user base upgrading and over what time frame that's reasonable?

Yes. The numbers we've given previously were more speculative and sort of had to do with what we thought might happen without many data points. And we now have some data points that look like they're falling about on that line. So I think this idea of 25% to 30%, call it, in the next year is probably about the right number. As I did mention, once we start shipping the 2500, then the price increase or to upgrade is going to go up, and it's going to go up pretty significantly. So we could have a situation where we get some -- a rush of upgrade orders before that price increase and then the upgrade slowdown afterwards.

Okay. And in terms of the pull-through. I know you mentioned on the prepared remarks how the pull-through on the 2500 would be greater than kind of what you're currently generating on the 2000. what's the kind of appropriate guidepost to think about on the 2500 pull-through on an annualized basis?

Well, that's pretty hard to know right now, I guess. Because there's -- people will use the 2500 in different ways. If they run it all out in rapid mode versus what they would do traditionally, you might think of it being about 20% higher. But that may not be how people use them because, particularly in the large centers, we'll have some mix of 2500s and 2000s. And so they may run all the fast stuff through the 2500 and then leave it idle and run the big stuff, the longer runs on the 2000 to make sure they're free if they get a hot sample and that they have to run. And so it's hard for us to figure out in advance what that equation might look like.

Okay, great. And maybe one more quick one. Just on your ovarian sequencing program when we're out -- Christian and Marc, earlier in the spring, we kind of asked what sort of timing for an update on that program. Just wanted to see if there's any more clarity on when you would expect to provide investment information on that program?

Yes. We think the next check in point of that will probably be in the fall.

Our next question comes from the line of Nandita Koshal with Barclays.

Jay, firstly, I was wondering if you've seen a meaningful step up in your services business associated with a faster turnaround due to the 2500?

Well, I'd say not yet, and a couple reasons for that. One, it's a little early since we announced it pretty recently. The second reason is that when we announced it, we announced a limited capacity because we didn't want to load up our services lab with a lot of 2500s until we understand what the market demand is. So I don't recall it exactly what the number is, but I think we announced 5 a week?

5 to 10 I think it was.

Yes, 5 to 10 a week is the maximum we could do. And so that's the most we would book. So it's not a high percentage in the overall samples. And we're going to use that here for the first couple of months. We get assessment of how big the market demand is on the RUO side. And we will launch this into CLIA, and then we'll start adding 2500s to track the market demand. So for those reasons, I think it's too early to really understand what the total demand might look like.

Okay. And is that your expectation that the CLIA part of it will really drive demand in that business, not really the RUO side?

I think the CLIA will become more important, yes.

Those are the types of genomes where there's a medical reason for it to be rushed through. And there are a great many cases per RUO genomes as well but I think there are fewer than the clinical cases.

Okay. Because it's going back to Ross' earlier point about how customers might be thinking about dealing with the uncertainty in the back half. Could they be thinking about service as a more viable option given the faster turnaround now, versus actually buying instruments and consumables? Would that be an option to sort of deal with?

Well, they could, but the faster turnaround cost more per genome. So the economics aren't in their favor there. I mean, it's variable cost, so they don't have to put out the capital. But it overall spends more money with -- per genome with Illumina. So I'd be surprised if too many customers make that decision. Some might.

Also, Jay, the product offering is a lot more narrow relative to what customers are doing. So on a rapid turnaround, it's really just whole genome sequencing and human only. So -- and if you look at what the majority of our customers are doing with their sequencers, it's everything from exome -- it's some whole genome. But I wouldn't say that's the dominant application right now. It's exome, RNA, small -- other small -- smaller types of projects are the key drivers of consumable pull.

Okay. Thanks, Christian. And if you could also comment on the competitive landscape that you're expecting in the second half in early 2013, obviously some new product launches coming up from competitors as well. Could you talk about those assumptions on you end? And then R&D priorities for Illumina, just what we can look forward to by the way of new product launches, apart from the ones that you already announced?

Well, I guess we haven't seen a lot of change in the competitive dynamics, maybe a little less competition on the services side. The products that have been preannounced aren't in the marketplace yet really, and so there's no real change in what we expected there. So we continue to believe we're going to be very competitive, particularly if you started looking at the MiSeq performing at the kind of levels that we talked about with these sample prep and ease-of-use and the BaseSpace attached to it. We think that's going to be a very competitive product in the overall market. In terms of new products, we obviously don't talk specifically about those. But we are working across the board, and you've seen us do a lot of innovative things in the informatics side. You'll continue to see that. We're clearly working very hard across the set of platforms that we have to enhance those. We're working on new generation technologies and various sorts. And we've had great success in our program to back-integrating the sample prep. And I think you'll continue to see us do that in a very aggressive way. And over time, the sample prep will become increasingly integrated with the system, particularly as we drive these products into clinical markets where ease-of-use and push-button operation is more and more important.

And indeed, it's not just one comment on the competitive landscape. I mean our current outlooks reflect the known competitive situation. So anything we get back from our sales team that sweep through the funnel reflects what they know about what we know about what's currently been announced.

Our next question comes from the line of Isaac Ro with Goldman Sachs.

I just want to, first off, try to address that question of visibility a little bit more directly. I'm wondering if you could simply say if book-to-bill was in fact greater or less than one this quarter?

Yes. We're not going to disclose it this quarter, Isaac. And it's -- and you can read anything into that. I mean it really is just that we don't want to get into a pattern where it's expected that we disclose book-to-bill every quarter. We've done it twice, and we don't want to be in the habit of feeling like we have to do that. We do report our backlog at the end of the year, so you'll see that in our K at the end of the year. And so you can make a year-to-year comparison about book-to-bill when we publish that.

Got you. Fair enough. And then secondly, on PCR, just wondering if you could maybe try and frame on the long-term kind of how you're going to measure success. On the one hand, the technology you have is obviously priced aggressively and applicable to a very large install base of systems. For the side, the incumbent is very well established. So if you could maybe put a timeline on how you how long you're looking at kind of measuring success and what those terms might look like?

Yes, I think we need to look at this as a long-term investment on our part. So we'll measure success over a period of a couple of years. It is going to take some changes in our distribution channel, which we've implemented internationally and have yet to do at the level we would like in the U.S. but I think we're close to getting that done. And so I think we're going to need to give that distribution channel some time to ramp up. It's not going to be as quick as our direct distribution channels typically would be for a new product. And so I think we'll measure success over probably the next 2 years. And we need to have good reach to sell these products because they're sold in high volumes into lots of labs and not typically the places that-- where all direct sales people would've called otherwise. And that's why the distribution model works well here. Hopefully, it will work well, as well as having lower ASPs where you can justify direct sales.

Terrific. One last question for me. It's a big picture one on MiSeq. It's been in the market almost a year now. And I'm just wondering if you could give us an updated view on how you frame the large -- the long-term market opportunity for that product cycle. And the reason I ask is I think most of us have tried to ballpark it. I think somewhere between the HiSeq market opportunity and then the PCR market. Obviously, that's a very wide range. So given the updated way that you're looking at the framework for the market there, that would be super helpful.

Well, I can say that we're very bullish on MiSeq and its future incarnations and derivatives of that platform. We've made great technological progress going from a HiSeq to a MiSeq and we're going to continue to drive all aspects of the chemistry in the optical performance of the system. So we think that if you think about the overall market opportunity here over some period of years, in this category of desktop instruments, it's certainly in the range of 10,000 plus whether it's 20,000 we could argue about, but we think it's certainly in that 10,000 range at a minimum.

Our next question comes from the line of Dan Leonard with Leerink Swann.

How are you thinking about the impact of foreign currency on your business in the second half of the year?

Dan, it's Marc here. We've -- obviously, we don't forecast what we think is going to happen with the FX rate. We do have a fairly natural hedge when you take into account where our cost structure is, particularly with the U.K.'s R&D and presence in the Singapore manufacturing. And when you look across our P&L, it nets out that's up pretty nicely, the currency is moving the direction they have been lately but we just don't know where it's going to go in the future. And we have a fairly tried-and-tested hedging strategy, which is pretty basic -- it's pretty simple. It's not overly sophisticated, but it's been working us lately. The effectively -- in the past, the last couple quarters have been material for us both to revenue and bottom line. So I can't tell you what it's going to look like going forward but we'll continue to track it the way we are and see if we can offset and mitigate some of the impacts.

Okay, thanks. And Marc, can you comment on the technology write-down, if that was in anyway related to chemistry A or chemistry B?

No. I mean, it wasn't. The technology write-down was an asset that we acquired a couple of years ago and it's not related to the DNA sequencing space.

Okay. And then my final question, what do you think your attach rate looks like in the sample prep business? So what proportion of your sequencing or incremental sequencing in cells are using your enrichment products as opposed to competitors?

I think our estimates now are somewhere in the 40% to 50% range of the client sample prep. A little hard to know exactly but we think that's the range and part of the problem is that there's some homebrew stuff going on, as well as stuff supplied by other competitors. And our goal clearly is to continue to take market share there. One thing I might add on the prior question as well is that, what happened with this technology write-down is that we go through every year multiple times when we look at our product portfolio and our research portfolio and decide how to prioritize products in that collection of opportunities. And in this particular technology, we decided to lower the priority enough that we decided that we needed to impair the asset because it has less clear visibility for getting to the market and that's the reason we impaired it.

Yes. Just to punctuate that, I mean, we're not abandoning that asset. We're just reprioritizing our focus.

Our next question comes from the line of Dan Arias with UBS.

I guess just getting into the MiSeq opportunity a bit further, Jay. Can you give us a sense for the level of the replacement that you're doing with the MiSeq at this point?

I don't actually have those numbers at my fingertips so I would say, anecdotally, maybe a little less than what we thought but our Amplicon product is starting to pick up steam, and that's really the market that -- or the sample prep methodology that goes after the biggest part of the CE market. So we continue to think that over time, those are going to get largely displaced in other than the sticky applied markets. But there are customers that are slow to change. They've got a process has been working for them for a while, and we need to just show them why this is so much more economical and so much easier to use than CE. So we're doing okay there but probably a little less than we might have anticipated.

Okay. And I guess are you able to talk a bit about the launch plan for the targeted panels that you're looking to offer for MiSeq? How often might customers be looking at new products and maybe what are the key focus areas there from that area to these perspective or genome region perspective?

Well, let me talk about that in general terms that we have quite a number of -- well, what we have a number of panels that we think we'll have in the market by the end of this year and those we marketed through our TCG group to existing CLIA laboratories for the most part, and these will be enabling panels that will allow these CLIA labs to go after specific disease areas. In some cases, they can add their own content to the panels but you think of them as enabling more our customers. And then the other kinds of panels are the ones that come out of our Greg's pure diagnostics business where there are actually products that will run our cells, and that's where we're putting the MiSeq through the FDA with our own panel in Illumina market directly.

Okay. And then maybe if I could sneak in just one more higher-level question. I guess as clinically usage gets going, how do you get to the industry standard for reporting and quality on the assays that will be run? Is that something that you think is a key hurdle and -- is there something that you can do there to help that process along?

You mean to actually achieve quality that's necessary or to convince people that the qualities were it needs to be or...

Well, just in terms of getting the customer base comfortable with the reproducibility of results and the standardization of what they're getting off from the machine?

Yes. Well, I guess the first thing is we're going through a number of clinical trials on MiSeq and that certainly is going to help as we publish the results of those clinical trials. It's going to certainly help with the ability to -- for publications in front of customers that show that reproducibility. We are putting them in front of the FDA as well so the FDA hopefully will put their stamp of approval on this system as analytically producing reliable and repeatable results, and then for the specific assays we've put on, and that those assays have clinical utility. We're working with the FDA across a broad front here to make sure that sequencing gets moved forward quickly, and I think the FDA is very on board with this idea and that they want to have sequencers approved and to demonstrate that these machines are reliable, that they do produce accurate results and in many cases, more accurate results than any of the existing technologies do. And I think as we get through that point at the end of the year, I think it's going to become pretty clear that these platforms are clinically ready.

Our next question comes from the line of Peter Lawson with Mizuho.

Jay, the strong service revenue, what's behind that? Is that whole genome sequencing or race? Could you break that down out in anyway?

That's a combination of, yes, whole genome sequencing which was good during the quarter, some array contracts and also a result of our growing install base and the services revenue that we get from the maintenance contracts.

And then across the customer base, are you seeing a resurgence in whole exome sequencing versus whole genome sequencing?

Not sure I'd characterize that as a resurgence. I mean, it certainly has been a strong, growing market segment and I think it continues to grow. If you look at the trade-off between exome and whole genome, you'll find people in both camps and it's largely an economics question and a question around the power of the type of study that they want to conduct. So if you're looking for certain characteristics or certain types of discoveries, you can do a whole lot more samples with exome, and that may be the better design if you want to a pure hypothesis free study and you want to look everywhere and the whole genome is the only way to go, but it will cost more per sample and therefore, your study will be less powered from a sample perspective. So I think over time, as we've said over and over that we think the exome market is a good one but will become less and less of a factor if you look out 2, 3, 4 years because the sample prep will become more a dominant cost on the front end for exome sequencing compared to whole genome and therefore, the economic gap will shrink.

And do you find at the moment, there's any difference in profitability between whole exome sequences and whole genome sequences in the sense of downtime and the difference in the front end prep?

In terms of our profitability?

Yes.

I don't think so. We've probably -- let's see, let me think about that. So in the exome case, we have a little more sample prep revenue because if they're using our exome prep kit, but the per sample cost on the sequencer is less because you can multiplex so many of the other side. We'd have to work that out to see what the relative profitability actually is between those. I don't -- I can't do that math in my head.

Our next question comes from the line of Amit Bhalla with Citi.

This is Adam in for Amit. I just had a question on the 2500 rollout strategy. How should we be thinking about that to the top account, and what's the latest details you can give us?

We expect to begin shipping in Q4 and we also expect to start putting out the upgrade kits probably around the end of Q4. So the upgrades to the install base. And as we mentioned, many of the -- or most of the incoming orders are for 2500s or 2000s with upgrade kits. So we'll start shipping those new units at the early part of Q4.

Okay. And then I know you've commented on your end market mix changing, becoming more commercial and that was lending better visibility for you. Can you quantitate that or quantify that more or give more detail on where those trends are going?

Yes. We're working hard to get a better database on this so that we can do a better job of knowing exactly where these systems are going. In some cases, for example in distributors sales, we don't have as good information as we might like. We've historically -- if you look back a year or 2, talked about it being about 80% in the academic markets and 20% into sort of commercial and nonacademic sorts of sites. We think that's probably down to 70% now, maybe trending a little bit lower than that but we're doing a little bit more analysis now to make sure we have those numbers right. Certainly, the incoming order rate is biasing more and more toward nonacademic accounts. So our overall exposure there is being reduced.

Our final question comes from the line of Zarak Khurshid with Wedbush.

Just stepping back a little bit, how much of your total consumable sales would you say are going into the clinical end markets today, and how fast would you say that stream is growing?

Let's see, what can I say? I would guess -- I don't have those numbers specifically so this is an estimate. If you want the actual number, you can call us back later and we'll do the math. But I would guess it's about 10% going into clinical markets, and we have a couple of very rapidly growing customers who have fully commercial tests in the market that now have a large install base of HiSeqs. If you look across the T21 testing market, there's 5 companies all using our platform. A number of these are commercially running tests every day now. So it's growing quite quickly and I'd say at a much faster rate than the research consumables are growing but I'm not sure we've actually calculated it in that sort of segment.

Got it. So would you say half that 10% is attributed to some of those prenatal companies?

Yes. I don't think we'd give a specific number on that, and I don't know what the number is so -- but all I can say that, that segment is starting to grow. I mean, the orders are clearly coming in there, and those companies are ramping.

Great. And just to clarify, the kind of lab question, Jay. So the breakdown of current MiSeq orders is, there's an overrepresentation of nonacademic accounts, is that fair?

That's right. Yes.

Our next question comes from Paul Knight with BLSA (sic) [CLSA]. Paul R. Knight - Credit Agricole Securities (USA) Inc., Research Division: Could you talk about your international exposure and when would we expect that tax rate to finally drop?

International exposure with respect to overall orders or were you thinking about with respect to the tax rate? Paul R. Knight - Credit Agricole Securities (USA) Inc., Research Division: Well, Jay, as PEAR as they have grown their Asian operations, for example, tax rates have drop, would you expect that?

Yes, we would. So Asia has been particularly strong for us over the past few quarters. Europe has been okay. Americas over the last 2 quarters at least is growing pretty nicely. So on a relative revenue basis, I'd say at least in Q2, the Americas was a higher proportion of the revenue. But certainly, as that mix shifts toward Asia, or if it shifts toward Asia, we would have an improved tax rate. Marc, you want to talk about sort of our expectation on how it will reduce over time?

Yes, Paul, I mean, we've -- as you've pointed out, we've been growing our operations in Asia and Singapore in particular. We've been moving manufacturing there at a reasonable rate and particularly are focused around our consumable manufacturing. And so that's helping us to improve our tax structure but we're going to continue down that path. We're looking at not only that but as obviously every other, in fact, opportunity that we might have, and -- but that's absolutely one of the levers we can pull and we'll continue to pull it in the future. I wouldn't want to try and give any guidance around what the tax rate might look like but I would expect to decline going forward. Paul R. Knight - Credit Agricole Securities (USA) Inc., Research Division: Yes, is this going to help your margin expansion? Or is it just implied?

I think, yes. I mean, a little bit, yes. Yes, I think it should.

And ladies and gentlemen, that does conclude the time that we have available for questions. I'd like to turn the call back over to the management for any closing remarks.

Thank you. As a reminder, a replay of this call will be available as a webcast in the Investor section of our website, as well as through the dial-in instructions contained in today's earnings release. Thank you for joining us today. This concludes our call, and we look forward to our next update in October following the close of the third quarter. Thank you.

Ladies and gentlemen, thank you for your participation in today's conference. That does conclude the presentation. You may disconnect. Have a wonderful day.