41J Blog

I never quite know where to place the BGI (originally Beijing Genomics Institute). Are they a sequencing service? Are the a instrument vendor? Are they a business at all? Or are they a research institute? At times they seem like all these things, and none of them. My clearest personal memory is from one of the Cold Spring Harbor conferences. The moderator announced that before our next talk a representative from the BGI would like to make a brief statement. The BGI representative stood up and read a pre-prepared statement announcing that the BGI would sequence 1000 plant an animal genomes… it was an ambitious project but clearly one that the BGI were capable of accomplishing (using Illumina machines at the time). But the delivery seemed weird. Usually these large projects are coordinated through an international consortium of researchers. The BGI just stood up and stated that they were going to do all this themselves… it kind of sounded like a declaration of war, and indicated to me that the BGI is something other than a traditional research institution.

At present they a massive fleet of Illumina machines, their own commercial instruments employing two different sequencing technologies, a sequencing as a service business, a clinical NIPT test, 1000s of researchers, and offices in 4 countries. So what exactly is the BGI anyway?

bgi.com in 1998

I figured I’d start by trying to go back to the beginning. The BGI was founded in 1997, and appears to have been developed out of China’s desire to take part in the human genome project. bgi.com wasn’t owned by the BGI in 1998. genomics.cn was most likely, but the earliest capture in the waybackmachine is from 2008, it’s quite sparse and talks of them having Illumina Genome Analyzers, ABI Solids and 454FLXs (it’s notable that only the Illumina range of instruments still exists). But it’s otherwise not very instructive.

The BGI history page, tells us a little more stating that “On July 14, 1999, BGI was founded with the mission of 1% of the human genome for the International Human Genome Project.”. In China research funding comes from the Ministry of Science and Technology . I would imagine funding came from the “973” program. Which was China’s basic research program until 2017, when it was supposed to be replaced by something else. 

Early versions of the Chinese wikipedia page on the BGI state that after the completion of the human genome project the BGI relocated to Hangzhou in exchange for local government funding, and then in 2007 they announced that they were to relocate to Shenzhen to establish China’s first private non-profit research institution, in 2008 BGI Shenzhen was approved by the Shenzhen Municipal Government to become a public institution. So from what I can tell they started off firmly as a non-profit research institution and remained so until at least 2008. What’s less clear to me is what kind of entity this was (non-profit? state owned? private company?).

2010 seems to have been the turning point for the BGI. They received a 1.58B USD line of credit from the China development bank [1], and funds from Shenzhen Capital Group [2]. It also looks like Taikang Life Insurance [2] and Sequoia may be an investors [4]. Then in 2013 after a long and seemly slightly painful process they acquired US DNA sequencing startup Complete Genomics [3]. Finally on the 14th July 2017 they were listed on the Shenzhen stock exchange (company 300676) completing their transition in to a somewhat surprisingly highly commercial research institution.

The Chinese wikipedia page now lists 6 business units: BGI Research Institute, BGI Technology Service Co., Ltd. , BGI Health, BGI Agriculture, BGI Genetics Institute and BGI Cloud Computing.

Glassdoor reviews for BGI Shenzhen seem pretty reasonable (mostly complaining about the pay). Glassdoor reviews for Complete Genomics post acquisition seem to indicate that much development work was transitioned to China, and tantalizingly that the “omega sequencer project” was killed.

That about wraps it up for the business side of things. I will continue to add information as I come across it (not sure what information Chinese companies need to make available, but if anyone knows I’d be most interested in revenue projections etc.). I’ve covered the technology they acquired from Complete Genomics and plan to cover other aspects of their business too.

Notes

[1] https://www.technologyreview.com/s/511051/inside-chinas-genome-factory/

[2] https://www.crunchbase.com/organization/bgi-2/funding_rounds/funding_rounds_list

[3] http://www.genomics.cn/en/news/show_news?nid=99460

[4] https://www.sequoiacap.com/china/en/companies/

Following on from my list of sequencing companies I wanted to delve into Base4 next. Below are my brief notes on their business and technology.

Business

If you’re not really interested in the business stuff, skip this and head to the tech section below. Suffice it to say they’re a series A/B stage DNA sequencing startup and have received about 21M GBP in total investment.

The company was incorporated in October 2007. Being a UK company there’s quite a bit of information at companies house [3].

Investors appear to include: Longwall Ventures, Meridian Corporate Finance, Oxford Technology Enterprise Capital, Royal Society Enterprise Fund, Torteval Investments Ltd, and Amadeus RSEF. Amadeus RSEF I think stands for Royal Society Enterprise Fund, and possibly indicate that this investment is associated with that fund, rather than the main Amadeus fund.

Companies house accounts list 36 staff in 2016 [4] , with a burn rate of about 3M GBP a year. They had around 1.6M GBP in cash. The accounts state “to meet the company’s ongoing cash needs…additional funding will be required…within the next 12 months”.  And their site states they received 5M GBP in November 2017 [1]. This may indicate that further fundraising would need to take place in 2018.

Glassdoor reviews [2] for the company contain some entertaining gossip. It looks like there has been some internal restructuring at the company, but with the raise last year hopefully things have settled down.

Technology

The image below shows the basic mechanism they present on their website. The essential idea is pretty straightforward. Take single molecules of DNA, use pyrophosphorolysis to pull off individual nucleotides. The individual nucleotides then get encapsulated in microdroplet reaction vessels where they undergo a reaction which allows then to be detected by a fluorescent signal. My guess would be that the idea is that this might result in longer reads as opposed to Illumina-SBS sequencing. It seems doubtful to me that this process might result in higher accuracy reads.In addition to this method they appear to have patents covering plasmon nanopores, which I’ll cover below. First I’ll cover each part of the process above in a little more detail.

First they need to remove individual nucleotides. The normal way of doing this would be to use an exonuclease. Which is an enzyme that cleaves off bases. Others have proposed using exonucleases for single molecule sequencing (including Oxford Nanopore’s original approach). The patents don’t very clearly state what they’re doing (when are patents ever clear?) but they appear to be using an enzymatic approach, I guess either an exonuclease or a polymerase with some exo activity. The sequencing template is attached to a surface and the nucleotides get chewed off.

They then get stocastically captured in droplets (so there will we a lot of empty droplets, and quite possibly some droplets containing multiple bases). These then get transported through a fluidics system (though a recent patent suggests they might also just be arrayed on a surface [5]).

At this point they have single nucleotides in little reaction vessels, presumably with some reagents. The next stage is therefore to determine which nucleotide is in each droplet. The website currently refers to a cascade reaction and fluorescence, but patents also refer to Plasmon resonance [6].

The cascade reaction appears to work by having the single nucleotide bind to a capture site on an oligo [7]. A substitution-dependent restriction endonuclease comes in and cuts the oligo depending on whether the nucleotide is present or not. The now cut oligo can now be processed by a double-stranded exonuclease. This releases more single bases, hence producing a cascade reaction.

In the process of the above cascade reaction, fluorescent labels are activated. The exact process isn’t described (that I can see) but it could be that a quencher is removed or something similar. Like most patent portfolios, this is just one suggested mechanism… but I’d guess this is the cascade reaction mentioned on the website. That about wraps it up, by detecting the fluorescence (I assume they need multiple labels/lasers) they can call the DNA sequence. There are also patents referring to methylation detection but I’ve not taken the time to read these yet.

As a final note, a few of their patents refer to plasmon nanopores. This appears to be unrelated to their microdroplet sequencing efforts, and for an early stage company it seemed odd to be experimenting with such a radically different approach. The patents contain SEM images, suggesting that some real work has been done on this system. The approach is obviously also somewhat related to the Armonica approach previously discussed. SEM images below. It seems likely that this may relate to their collaboration wit Hitachi that was reported in 2013 [9].

That about wraps it up for Base4. There’s some interesting tech, and a pile of patents which may no doubt yield further insights. If you’d like to discuss Base4 further (or sequencing in general) leave a comment below or email me (new at sgenomics dot org).

Notes

[1] http://www.base4.co.uk/news/base4-nov17-funding/

[2] Glassdoor reviews (selected mostly negative review containing interesting gossip):

[9] https://www.genomeweb.com/sequencing/base4-hitachi-developing-single-molecule-nanopore-based-sequencer

“Privately owned Base4, the partners said, has developed a method of increasing the signal from the single molecule passing through the pore by using laser light enhanced by gold structures. The technology, they added, allows the signal to be read from unlabeled DNA, minimizing sample preparation.”

In my last post I put together a list of 41 companies which either have a DNA sequencing platform on the market, or in development. There is really only one company that dominates this market (Illumina) so most of the players are either niche plays or startups. Armonica is one such early stage startup.

Business

They appear to have raised 1.5MUSD [1]. Before looking into the technology, there are a number of things that stand out about Armonica. Firstly it’s an LLC, typically the received wisdom for creating a venture funded tech startup is to create a Delaware C-Corp. From what I can tell, creating an LLC could be problematic for traditional investors (who might want preferred shares for example) [2] (appears to now be listed as Armonica Technologies Inc. [13]).

The next thing that stands out is the list of “Board Of Directors/Executive Leadership”.

There are 5 people on this list [3], but of them only Steven Brueck (CSO) and Scott Goldman (CEO) appear to be part of the company (the rest are investors). Sally Corning is listed as co-founder, and a partner at a fund, which seemed odd to me.

A LinkedIn search shows Scott and one other senior consultant. The CSO is a university professor so most likely doesn’t do much at the company day-to-day. An SBIR entry suggests they have 23 employees which seems surprisingly high [12].

Overall, it seems to be seed stage, the company was founded in 2015 [4].

Technology

Armonica is an optical nanopore DNA sequencing company. Optical nanopore approaches use a nanopore to control the motion of a strand of DNA but an optical readout (as opposed to direct electrical detection [5]). Armonica are not the only company to propose the use of an optical nanopore, Quantapore have presented a protein nanopore, FRET based concept too. But the method suggested in Dr Brueck’s patents [6] and paper [7] differs significantly. Firstly, the fabrication methodology they’re suggesting is quite novel, and secondly the optical readout is label-free (using surface plasmonic resonance).

First up, fabrication. There are a number of novel features here firstly they’re using interference lithography to define their features. Interference lithography is one technique that lets you beat the diffraction limit and create features smaller than the wavelength of light. It does this by using the interference between two coherent light sources to define the pattern. Pretty neat! The downside is of course that you are limited to periodic structures, but that seems fine for nanochannels and they present SEM images with channels as small as 100nm.

After using interference lithography to lay down photoresist they then lay down Silica nanobeads, after which they heat the whole thing, baking out the photoresist and sintering the silica to create the final structure. Neat! The exact advantage of using nanobeads over other deposition methods is less clear to me (I’m no expert) and some quick googling brings up some alternative methods for creating enclosed nanochannels [8]. It would be interesting to dig further into this aspect of the technology, and contrast the various approaches here.

With nanochannels in place, they can DNA through them (see below) . This is where their paper finishes up. The Silica is transparent, so they can image YOYO stained double stranded DNA right through it. They note that in the small (100nm) channels, it’s harder to get the DNA moving through, and that it moves more slowly (driven only by capillary action here).

DNA imaged in nanochannels from [7].

The above is based on published work out of Brueck’s lab. The patents (and brief description on the website) mix things up a bit, and describe a tortuous nanopore. In the patent tortuous nanopores are rather aptly described as “complex, 3D paths, similar to the spacings and open paths created when oranges are piled up in the local supermarket”. So, the idea is rather than precisely defining a nanopore, you can stochastically define them with these piles of silica nanobeads. You can also use a more traditional deposition method to (somewhat selectively) close apertures in this “tortuous nanopore”. This also seems like a rather neat idea. The tortuous nanopore concept potentially provides a method for slowing translocation (generally an issue with solid state nanopore systems) and for generating small apertures than available with traditional fabrication processes. It makes me wonder if the system couldn’t be used for ionic current detection (or perhaps there is too much stray ionic current, or other issues).  From what I can tell the system uses a bias voltage to drive the strand through the pore.

The base detection process is probably the least detailed part of the patent. My understanding is it essentially states that they’ll use surface plasmon resonance, and that there are a bunch of ways of doing that. Using SPR for DNA sequencing isn’t a new idea, and a quick google brings up grants for as early as 1999 [9]. But I would have to do more reading to know how practical it is here. Interestingly another company (Redondo Optics [10]) has an SBIR grant [11] for using SECAR SPR for DNA sequencing. The CEO of Redondo optics is also on the Armonica SAB. With more time this would be an interesting avenue to follow up.

For the moment however, I leave this brief review here. Overall the Armonica technology seems to have a number of interesting novel aspects. If you have any further insights into the technology, or would help to chat about DNA sequencing drop me a line (new at sgenomics dot org), or leave a comment!

 

[1] https://www.abqjournal.com/1050880/unm-technology-accelerates-genomic-sequencing.html

[2] https://mashable.com/2016/01/27/startup-business-structure/#eSxUGPb_qGqX

[3] Currently listed on the website are:

David Blivin, MBA Managing Partner, Cottonwood Technology Funds

Steven R. J. Brueck, Ph.D. Chief Scientific Officer, Armonica Technologies

Sally Corning, MBA Co-Founder & Partner, Sun Mountain Capital

Scott Goldman, MBA President & Chief Executive Officer, Armonica Technologies

Waneta Tuttle, MBA, Ph.D. Fund Manager, Tramway Venture Partners

[4] SEC Filing https://www.sec.gov/Archives/edgar/data/1714729/000171472917000001/xslFormDX01/primary_doc.xml

[5] Most other nanopore companies propose detection through the blockage of ionic currents, or tunneling currents.

[6] US9927397: https://patentimages.storage.googleapis.com/56/92/ad/2b5ef43d1f4c47/US9927397.pdf

US20160377590A1: https://patentimages.storage.googleapis.com/d1/8a/16/0bc8e6aafcf270/US20160377590A1.pdf

[7] http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.172.7987&rep=rep1&type=pdf http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.172.7974&rep=rep1&type=pdf

[8] https://www.princeton.edu/~chouweb/publications/148%20Austin_Scanning%20the%20Controls_IEEE%20Nanotech_2002.pdf

[9] http://grantome.com/grant/NIH/R21-HG002102-01

[10] http://www.redondooptics.com/

[11] https://www.sbir.gov/sbirsearch/detail/1031029

[12] https://www.sbir.gov/sbirsearch/detail/1490917

[13] https://www.sec.gov/Archives/edgar/data/1714729/000171472917000001/xslFormDX01/primary_doc.xml

Below is a list of 40+ companies who either have a DNA sequencing platform on the market, or where there’s public information that they are developing a DNA sequencing/mapping platform. Below the main table is a list is mapping companies, many of which have sequencing related IP/approaches. There are a few cases where it’s not clear if the company is still active, I’ve tried to note this. If have any information on these companies, know of any others, or would just like to chat about sequencing companies/the sequencing market in general, feel free to email me (new at sgenomics dot org).

Disclosure: I’ve worked at a few of these companies, all the information presented here is based on public information.

Due to the width of the table, it’s best viewed on desktop.

Mapping

SBS: Sequencing-by-Synthesis

SBH: Sequencing-by-Hybridization

SBB: Sequencing-by-Binding