Wednesday, 14 January 2015

Illumina's announcments from JP Morgan

For those that are interested you can listen to the presentation by Jay Flatley at JP Morgan. And if you can't be bothered, then here are my notes (be warned they may contain inaccuracies):


The last 12 months: 2014 was massive for Illumina - X Ten predictions of shifting five systems were surpassed with 18 customers and 201 units, NextSeq was a big performer, and there was a distinct lack of MiSeq and HiSeq install cannibalisation. Revenue of $512M in 2014 Q4 and an expected $1.86B for 2014, Illumina expect 20% growth in 2015 and reported margins of 73% (hmm I really should blog about that sometime).

NextSeq: Over 500 NextSeq instruments have been installed and Illumina expect $80-100k reagent pull through per unit (wow - why do people buy these expensive instruments when there is almost certainly a lab willing to share less than 10 miles away!). The 2nd gen NextSeq chemistry with improved error rates should arrive in the first-half of this year and is now consistent with HiSeq/MiSeq. The NextSeq will be locked down for regulatory approval in late 2015/2016 ready for NextSeq DX. Jay also introduced the NextSeq 550 which allows microarray scanning with minor upgrade, Illumina's focus is on cytogenetics (no gene expression arrays at launch - RNA-seq kills HT12?) and the 550 will generate the same data as iScan. Costs is $275k, $25k premium over 500, upgrade available for just $50k, shipping in Q2.

HiSeq X Ten: $1000 price point has had a psychological impact (but is anyone really getting $1000 per genome?) Illumina are seeing $600-650k reagent pull through per instrument. Genomics England is best example to date of application of genomics to health care (awesome to see that Illumina are looking to the UK), aims to prove economic benefit to patients/NHS. Illumina have currently run 2000 rare disease samples. The X Ten install base continues to grow. But the "brilliantly called" X Five should allow it to grow faster. Illumina can manufacture HiSeq X at any scale required (get ready for X 1?)
Costs are $6M for X Five +$1.2M per unit to 9. Run costs are just over $10Gb or $1400 genomes on X Five (HiSeq X Ten $7 per Gb).

Turnkey sample prep: NeoPrep had some technical challenges and success rates delayed launch. Now in beta test planning on shipping this quarter. About half of library prep catalogue should be available. 10-50ng inputs reduce nucleic acid requirements. (Blog topic: Should you buy a NeoPrep?) 
     HLA market $250M for transplant etc, 110 MiSeqs into this market - an integrated solution (MiSeq plus NeoPrep) should allow this to be taken up in many of the 450 HLA labs. Illumina's HLA assay covers 11 genes giving 10x more information than today with full gene phasing.
     Forensics $400M today - STR has big limitations. Integrated method is backward compatible with CODIS database, sequence 200 regions ancestry, ethnicity and physical traits. 10 systems ordered.

BaseSpace: the worlds largest genomics cloud, over 3000 sequencers and 125,000 runs. Proteomic apps, NextBio Apps, BaseSpace labs by Illumina scientists (and MGA one day too). Over 4500 users in BaseSpace and 30% doing some analysis. Log data from 8000 runs per month and aim to let customers get access to this data (can't wait to get a look at this).

NextBio: aiming for seamless sample to patient analysis, knowledge base continues to expand: 2500 papers, 27k patient records and 30,00 new variants. RNA-seq body atlas 50 Human tissues normalised. 2015 add capabilities to allow population genomics analysis.

NIPT: Illumina bought Verinata two years ago. Sex chromosome, microdel, trisomy 9  and 16 in Veriseq PGS. NIPT move from high-risk to average risk patients. This is going to grow and grow. Strategy to give foundational tools, no sale team. Chinese market is huge (but how much will be to choose sex!) Illumina processed 43,000 samples in Q4 2014, from labs sending samples in for testing, ther are over 50 labs globally collecting samples. Sequenom acquisition. Veriseq NIPT assay NGS HT 48 sample NextSeq run PE-seq, foetal fraction QC, no PCR, available in Q2 and CE marked in EU. Suspended efforts to submit HiSeq to FDA for NIPT. NextSeq DX in late 2015/early 2016.

ctDNA "the most rapidly growing diagnostic opportunity", the best NGS in the market not in retail business (not selling tests). The Actionable Genome Consortium (AGC), includes  CRUK - expect publication in a couple of months. Cat1 actionable today, Cat2 relevant but not yet actionable. Report Cat1 to oncologist, if no treatment options then Cat 2 reported and clinical trials suggested.
3x OncoPanel partnerships set up - aim to develop companion diagnostics via these.
Most exciting breakthrough is ctDNA in blood "the liquid biopsy". Therapy selection. Monitoring MRD in solid tumours. Discovery of new mutations to treat emerging clones. Holy grail is a screening test. This market opportunity is enormous optimising assay for detecting rare ctDNA in blood. SNVs gene fusions and amplifications. Sensitivity is incredibly important. Partnering with academic labs.
Aim to have a RUO kit by the end of 2016. Present clinical utility data (CMDL). Provide technology rather than offer lab tests. Solid tumours sequenced and mutations found, ctDNA assayed and always found the same variants, and also more variants than in solid tumour. detect cancer earlier, guide therapy choice, give feedback on therapy effectiveness.

HiSeq 4000: $900,000 New configuration using patterned flow-cells (at last) 1.5Tb or 5Billion reads in 3.5 days, $20 per Gb, no limits on applications, shipping this quarter. Demo data on BaseSpace (BLOG). Users need to get on the patterned flowcell roadmap. HiSeq 2500 drops to $690k. HiSeq family grows (but is this the end for 2500?)
Illumina expect customers to have multiple versions of the instruments (how to pay for it!?)





1 comment:

  1. What a classical research, you might have spent quality time doing this analysis or to summarize the whole presentation in the form of this article. Amazing work thanks for sharing.

    ReplyDelete