This month’s “Company Spotlight” takes a closer look at Blueprint Genetics, a rare disease diagnostics company on a mission to deliver the highest quality standards for genetic diagnostics.
Tero-Pekka Alastalo, Chief Medical Officer of Blueprint Genetics discussed with us in detail Blueprint Genetics’ offerings, its approach to developing high quality diagnostics standards, and its vision to connect physicians across the globe to help them identify and address unique rare disease cases faster.
Blueprint Genetics is a company that started with an intellectual property, the OS-Seq™ Technology, developed at Stanford by a Finnish researcher. The company thus is bi-continental with headquarters both in the US (San Francisco) and in Finland (Helsinki). The 150-person company was founded in 2012 by Juha Koskenvuo, Samuel Myllykangas, Tero-Pekka Alastalo, and Tommi Lehtonen. Blueprint to date has secured $26.3M in funding which will be used to get closer to Blueprint Genetics’ vision to bring the use of genetic information to mainstream healthcare with uncompromised high quality, by elevating the sequencing technology and clinical data interpretation platform to the next level with increased efficiency and level of automation.
The following summarizes questions and answers from my dialogue with Tero-Pekka Alastalo.
EB: Tell us more about Blueprint Genetics. What is the history of Blueprint Genetics from when you started to today?
Tero-Pekka Alastalo: Originally Blueprint Genetics was a Stanford University spin-off. The company was founded around an innovative new Stanford DNA sequencing technology, the OS-Seq™ Technology (Nature Biotechnol., 2011), developed by a Finnish post-doc (Samuel Myllykangas). I myself am a medical doctor by training with expertise in pediatrics and pediatric cardiology. We got very excited about this innovative targeted sequencing technology and its opportunities in diagnostics. Since our initial career plans were to return back to Finland to continue our clinical and academic careers, we founded Blueprint Genetics in Finland in 2012. With Stanford being the place where the technology was developed and owned, we knew from day one that Stanford will always be a very close partner.
In the early days, running diagnostics tests in hospitals, was known to be a long, time-consuming process with extensive turnaround times and high cost. We realized the potential of new technologies and innovations in DNA sequencing for genetic diagnostics, in particular due to their rapid evolution. At the time we founded the company, many of the clinical diagnostics laboratories were not even applying next generation sequencing, so they were predominantly relying on older technologies which have their obvious limitations.
We saw an opportunity to start pushing a new standard for genetic diagnostics. Speeding up the whole process, providing very high-end quality comprehensive sequencing, and offering it all at a lower price, to make it more accessible. But before moving it all into place and being able to provide the new technology on the market, we spent a lot of our efforts on R&D.
From the beginning the focus has been on human rare diseases. There are some hereditary cancers that are also of interest in this area, but it’s not really one of our major focus/offering.
Blueprint Genetics has by now been in the business for over six years and since our mission and philosophy has continuously evolved. Along the way, there’s been an evolution of the various technologies in parallel to us constantly developing new tools and improving them to elevate the standard of quality and performance for rare disease diagnostics.
EB: Next-generation sequencing (NGS) technologies have revolutionized genomics. How have they contributed to the understanding of rare diseases and diagnostics?
Tero-Pekka: NGS has significantly expedited our understanding of human monogenic disorders. I would like to say that there are thousands of diseases we now have some knowledge about at the molecular level. NGS has the power of saving cost, and in parallel enables the accumulation of many therapeutic applications that help uncover the exact molecular genetic cause for many diseases. This is of great interest, particularly for rare diseases where an understanding of possible mechanism can effectively accelerate orphan drug development thus impacting at a higher level the field of genomic medicine.
EB: What scientific and business needs is Blueprint Genetics addressing and what products/services are you offering?
Tero-Pekka: Genetic diagnostics as we see it today, is still a very young area with many novel next-generation technologies on the horizon that are not yet even on the market, or may not be in a diagnostic setting for a long time. Today, there is still limited understanding of how these technologies really work and perform, and whether they can be applied to diagnostics. If you look at all the hundreds and potentially thousands of laboratories globally who are trying to use these technologies in a diagnostic setting, it feels more like a jungle out there. There are so many different ways people and laboratories have decided to use various technologies for different applications and strategies. Standardizing these technologies for diagnostics purpose is a basic requirement – a need that Blueprint Genetics aims to address with its product offerings.
Furthermore, many laboratories are calculating the cost of sequencing and are looking into cost savings by changing the chemistry, or changing the way the machines are running. With rare diseases, we’re really looking for a needle in a hay stack. Laboratories are providing next-generation genetic testing, but at the moment customers who are sending respective samples to these laboratories don’t have an understanding what is happening with their samples, what the quality or the test performance of the provided tests is, how well the tests are performing for the genes selected, how much effort and money these laboratories are putting into running the tests, and other specifics around the laboratories’ processes. As of now there is a lack of transparency in the field of diagnostics. For Blueprint Genetics it is all about changing this and addressing these various components with full transparency!
EB: At this point do you still use your proprietary OS-Seq technology for sequencing your samples, or have you since changed the platform?
Tero-Pekka: In 2017 we started setting up a certain quality standard and comparing and validating existing whole exome sequencing platform chemistries. We identified one particular chemistry that worked very well and as a result we tailored and customized this chemistry. We launched in 2018 with this new whole exome sequencing platform which is based on a commercial chemistry and is run on the Illumina NovaSeq Sequencing System. It’s a different platform from when we started Blueprint Genetics – we transitioned away from the original OS-Seq Technology-based platform in the beginning of 2018.
Our new platform is a very unique exome sequencing platform. In our diagnostics setting, we’re seeing an over 200x mean-coverage with 99.5% of the target nucleotides having a coverage of 20x or higher. Hence, our new platform is a very high quality exome platform that has been tailored to cover non-coding variants. And as it has become clear, a lot of human diseases are associated with non-coding variants and a lot of hard to sequence genes. Our platform has provided a substantial boost to our ability to interrogate these non-coding and difficult to sequence regions at a high quality.
We are now running our exome platform for all of our samples and we can cater to specific customer requests. We can customize easily and therefore can do single genes, panels, or whole exomes. We recently published a comprehensive white paper describing the platform validation demonstrating that our platform outperforms every other offering when it comes to complex variations such as some very difficult-to-sequence, clinically important genes, indel, and copy number variant (CNV) detection.
EB: Data and quality transparency is an important aspect for Blueprint Genetics. What is it exactly that you communicate and make transparent when sharing results with the patient, physician, or a laboratory? Why is quality such an important component and how do you achieve it?
Tero-Pekka: Delivering results at the highest quality is and should be one of the major goals for companies like ours. Transparency has many faces and of course, many different aspects. I think one of the most important aspects is demonstrating the quality of performance or analytical validation of the platform. Currently, there exists no industry standard to provide this kind of information open to the public. Other genetic testing companies typically show zero validation data for their whole exome platform.
At Blueprint Genetics we use publicly available data and sample sets, which are fully traceable. This allows anybody to replicate what we do. And of course, we look at the validation from different angles – we are not just looking at sensitivity and specificity but are also looking at operational performance metrics such as how reproducible and how repeatable the platform is, among other aspects. The analytical validation is done with a large data set. We look at single nucleotide variants, different varieties of indels, and chromosomal defects, like microdeletion syndromes, to name just a few. Basically, we are demonstrating that our platform works with different types of mutations. Furthermore, this validation also provides a good understanding what the platform limitations are. All this data is been made publicly available via the previously mentioned white paper.
Another aspect of transparency is communicating quality metrics in addition to testing results via our online portal which can be accessed by health care providers. They can see:
- The comprehensive clinical report
- The quality and performance metrics for the analysis, and gene level information on performance.
EB: How do you differentiate yourself from your competition?
Tero-Pekka: There is very strong competition in the market when it comes to pricing and turnaround time. There’s a big risk that patients are diagnosed with approaches that are more cost-driven. The various laboratories are trying to survive in this market, I mean financially. We at Blueprint Genetics are fighting this trend, as for us it is all about the sequencing quality, including the quality of bioinformatics and interpretation components. We are transparent in every aspect and show the customer what happens with the samples and testing results and what the performances are for a particular test. As I said, along with our results we present analytical validation and how well our platform performs for different mutation types. We are sharing the platform limitations while demonstrating its performance. Most of our competitors are providing test results in a non-transparent way.
Therefore, one big differentiator relative to our competitors is definitely the way we’ve structured our platform and set up those quality standards. With these standards we can boost the diagnostic yield in rare disease diagnostics. That’s one of the key features we’re trying to teach our customers who currently might be going for a cheaper test that comes with significantly lower quality and performance metrics which consequently has a risk for lower diagnostic yield.
EB: Who do you view as your current competition?
Tero-Pekka: In North America, some of the major competing companies are commercial laboratories that include Invitae, GeneDX, and PreventionGenetics among other companies that provide rare disease genetic testing. If we look at some of the other territories around the globe, like Europe, we would say that the local university laboratories are competitors for us. However, they often struggle in achieving similar quality than commercial laboratories, as their diagnostic volumes are significantly lower and they have less capabilities in keeping the technologies up to date in this rapidly evolving field. These laboratories sustain their business often through political decisions in restricting commercial laboratories to offer testing. I have seen many examples where the offered local diagnostics is so outdated that many patients are left without a diagnosis that could have been achieved with latest the technologies. To summarize, our competitive landscape is composed of a wide spectrum of other commercial laboratories as well as different types of university-run laboratories.
EB: Blueprint Genetics is providing services for the entire workflow from samples to analyzed data delivered via a report, is that correct? If I were to have already samples processed, can I utilize your analytical services only?
Tero-Pekka: We have a full service laboratory and that is basically our business model. We provide every component. Yet, we are currently evaluating the capabilities of providing only the analysis services as a standalone offering because of our years of experience of establishing our bioinformatics solutions and also based on the strengths of our unique clinical interpretation platform. I would say, in the near future it will be possible that we will offer to our clients the analysis and the clinical interpretation of submitted sequence data as a standalone solution. This way we will be able to address different components based on different customer needs.
EB: Can you explain some of the specifics you apply to the interpretation of your data. What kind of algorithm(s) are you using and for what purpose?
Tero-Pekka: Our clinical interpretation platform has been built internally by our own engineers and is based on our everyday experience in analyzing rare disease patient data. This proprietary platform utilizes all the previously identified knowledge in Blueprint Genetics, all the previous interpretations, all the previous reports, etc. It also accumulates data from all existing databases. It’s a platform that curates variants and pulls out evidence for different types of variants from different databases. So, it’s really trying to make the interpretation process more reproducible, more reliable, and as such more accurate.
In addition, we have a team inside Blueprint that is building artificial intelligence based systems for variant curation. Our goal is to automate the entire interpretation process more and more. This is important as interpretation is the expensive part of the diagnostics process for rare diseases. For example, some patient data even today may still require up to a full workday of a geneticist to correctly identify the diagnosis and write the report. Compared to simple diagnostics, rare disease diagnostics can require a lot of additional time and effort. This needs to be optimized and made more cost-effective. I think this is where we really need this type of computational interpretation assistance enhanced with automation and artificial intelligence tools. We are currently very actively focusing all of our efforts on these above mentioned elements – another component where we differentiate ourselves from other companies in this sector that are developing these kind of platforms. Often, IT companies with these types of platforms, use existing databases or general knowledge to interpret their data. We at Blueprint Genetics use a machine learning process that is enhanced by every real case and includes the experiences of tens of thousands of patients. But of course, our platform also integrates publicly available information and content from existing public databases, such as ClinVar and OMIM, and databases accessible via open links.
EB: Is there truly a machine learning component that you apply or would you say you are comparing results to existing data?
Tero-Pekka: Yes, as I just mentioned, there is a machine learning component which we’re building inside our platform. This is in addition to standard comparisons to tens of thousands of patient data and using standard, existing databases. Many of these new tools are currently in prototype use. Eventually, we are planning to apply them to real-time diagnostics.
We are also trying to automate as much of the processes as possible. At this moment, still every report is written by a human geneticist and hence there’s still a lot of hands-on work that is required. Human report handling is not going away for a while. We definitely need it at the moment. Clinical interpretation is really a paper intensive process. And, it’s maybe the most expensive component of rare disease diagnostics. The end goal is to build tools that allow our geneticist to produce more reports and more statements within the same timeframe. But we need to make sure that the final reports will be identical in execution and quality to the ones generated purely by geneticists.
EB: How many geneticists currently work at Blueprint Genetics that are finalizing the clinical reports before they are delivered to the test ordering party?
Tero-Pekka: The headcount for fully qualified, full-time PhD geneticist is somewhere between 25 and 30. We also have a crew of clinical consultants who participate in the report accreditation process. We might be looking at about 50 people in total that are somehow involved in the results interpretation and final report delivery.
EB: Who would you say is your target audience for your clinical reports? Is it mostly hospitals or clinical testing labs? Do you have users/customers across all continents or is it mostly the US and Europe that you are catering to?
Tero-Pekka: The largest target audience of our products is best described as the large university hospitals with genetic departments. Having said this, we get samples from all kinds of clinics and hospitals. One common denominator of those ordering our tests is that they are physicians specializing in rare disease work, and therefore see the most sick rare disease patients. This explains why university hospitals may be a little over-represented when it comes to rare disease diagnostic testing and research.
We are at the moment operating on all continents. Of course, we have a big presence in North America, which is our largest territory when it comes to sample volume and number of customers. The second largest territory is Europe with customers from Nordic countries all the way down to the Mediterranean. All of these territories are growing very fast. In addition, we have an operation going in the Middle East and we have also initiated operations in Australia and in South America.
EB: In one of your recent press releases you mentioned that your new platform will be able to connect clinicians based on matching rare variants? This is a very interesting concept and the value is undisputable. How do you ensure patient privacy in the context of this capability?
Tero-Pekka: We of course ensure total patient privacy at all times. The test ordering clinician can only see that there is a match at the variant/gene level somewhere else in the community, and can request to be connected to that physician via the Nucleus database. The way it works is that the physician of the patient with the “matching variant” will get an alert that there is another physician in the Nucleus database who has an identical case. Technically, it is a portal that opens the doors for the two physicians to connect, but both physicians have to accept the connection. The rest happens pretty much outside of Blueprint Genetics. We are only helping to match physicians who have identical patients. It is up to the physicians and their patients how much they want to share among each other. Blueprint Genetics is not sharing any patient information, we are simply creating opportunities.
EB: Were you already able to connect some of those patient cases with physicians?
Tero-Pekka: Oh, yes! There are already hundreds of successful connections made. We don’t quite know how many of those connections resulted in further interactions and possible collaborations, but we’ve been getting increasingly positive feedback about this Nucleus database feature. It’s particularly interesting for physicians who are dealing with rare disease cases. Through our facilitated connection they now have someone else to talk to about their respective cases, and can discuss treatment options.
Currently the platform is a prototype (see the example screenshot below, a mockup, which visualizes what clinicians see when using the “Nucleus Connecting Clinicians” services), but we see lots of opportunities to develop it further. The platform even allows individuals who have not been diagnosed by Blueprint Genetics to search the database for matches. If there is a match a similar alert is sent to the physician with the match. We want to make it available to everyone.
EB: At this point how many tests have you performed and how many variants are in your system? What are some interesting numbers you can share with the audience?
Tero-Pekka: At this moment with our recently implemented proprietary technology and the new platform we’ve processed over 30,000 patient samples. All of these cases are rare human disease samples, covering about several hundred to perhaps a thousand different rare diseases (and growing) which overall represents a respectable database.
EB: Where do you see the field of genetic testing/clinical genomics move towards? Once we have overcome the whole data sharing challenge, how important a component will artificial intelligence (AI) be in the entire process?
Tero-Pekka: Eventually, we will be able to provide at a relatively low price, high quality sequencing-based testing. Although at the moment, I would say the laboratories are still providing data on a broad sequencing quality spectrum. Eventually, sequencing will be considered a commodity and therefore, sequencing will not be the biggest bottleneck. As the prices go down it will be relatively easy and simple to get comprehensive sequencing data, such as genome and exome data.
I believe the biggest pressure will be on the interpretation side. How do we interpret that data? The interpretation is just getting harder and more complex, mostly because we understand more and more about our genome, and we understand more about the various diseases. Interpretation is the biggest challenge we are facing and as a result this will be the biggest differentiator across the many laboratories that offer or will offer these types of services, and one of the main reasons why we at Blueprint Genetics are putting such a big effort in developing our interpretation capabilities. In the near future, as many individuals will have their genome sequenced at high quality, they will require high quality interpretation. As we learn more and add new disease knowledge repeated interpretation of those genomes will be the norm. I think that’s one of the areas which will be very fascinating.
The utilization of genetic diagnostics at the moment, is really kind of in line with the overall healthcare system. We are currently diagnosing and treating patients who are typically already in a very advanced stage of disease manifestation. This is what diagnostics is today! Most samples we are processing today are from patients that are very symptomatic. But eventually we can use this data to detect various diseases at a much earlier stage, or even predict these diseases and start early preventive actions for some of these diseases. I would like to use genetic information to predict diseases and try to improve the outcome for some of these patients, more expediently and more guided by following/applying preventive actions. But we’re still quite far from that. A lot of research still needs to be done! Though, I believe that’s essentially where I would see the direction of genetic disease diagnostics heading towards.
Also, I believe with the concepts of prediction and prevention, we’re actually talking about the impact on the broader human population. The current challenge we are facing: is how do we manage large populations with various genetic risks as the health care system is built for disease management and not disease prevention? As a result we’re facing a massive volume of individuals that require consultation. Using genetic information for prevention is a side product of a well-established genetic diagnostics process, but we’re not there yet. Currently, it’s much easier to provide diagnostics for an affected, sick patient than making meaningful predictions for a healthy individual. To be successful in a large scale setting the price of genetic screening has to significantly come down and we need to understand more about the genetic variants such as their penetrance for manifesting disease, and also how to manage the population with known genetic risks.