Name: Ralph Oehlmann
Position: Director of business development, IMGM Laboratories, Martinsried, Germany, since 2005
Experience and Education:
Field application specialist (several genomic platforms), Applied Biosystems, Darmstadt, 1996-2005
Scientist, Bernhard Nocht Institute for Tropical Medicine, Hamburg, 1994-1996
Scientist, Institute for Human Genetics, Medical University of Lübeck, 1993-1994
Postdoctoral fellow, Thomas Jefferson University, Philadelphia, 1991-1993
Clinical research assistant, Kabi Pharmacia, Erlangen, 1990-1991
PhD in biology, German Cancer Research Center, Heidelberg, 1990
Undergraduate degree in biology, Heidelberg University, 1986
IMGM Laboratories was founded in 2001 as a spinoff from the Center for Human Genetics and Laboratory Medicine Dr. Klein and Dr. Rost, a privately owned clinical diagnostic laboratory located in Martinsried near Munich.
The company, which has about 15 employees but shares additional staff members and resources with its parent firm, currently offers a variety of genomic services, including next-generation sequencing on the Roche/454 GS FLX and GS Junior platforms. It is also in the process of implementing the Ion Torrent PGM and the Illumina MiSeq. IMGM's customers are divided between pharmaceutical companies, biotechnology firms, and academic researchers, both in Germany and abroad.
In Sequence recently visited IMGM Laboratories, which shares a building with its parent company in Martinsried, and spoke with Ralph Oehlmann, the firm's director of business development. Below is an edited version of the conversation.
An accompanying video of IMGM's facility is available on GenomeWeb's video channel here.
What genomic services does IMGM Laboratories provide?
Our portfolio covers RNA services, which has been the main focus so far, including analyzing gene expression by microarrays or qPCR; DNA services, which is a fairly broad collection of technologies, from SNP analysis to array CGH to sequencing; and bioinformatics. The fourth field is consulting, for example for customers who come to us because they are developing a prototype of some sort and would like us to get involved to help with the development.
How do you differentiate yourselves from other genomic service providers?
Something that really stands out is that we work in close collaboration with our diagnostics partner, the Center for Human Genetics and Laboratory Medicine Dr. Klein and Dr. Rost, which means that we can bundle very interesting applications that can normally not be found in the portfolio of a general service provider, for example HLA typing, or flow cytometry, or karyotyping.
Also, sequencing is just one part of our portfolio. So we could identify interesting genes or biomarkers using microarrays, validate that using qPCR on a larger sample size, and then if it is of interest to a customer, we could go a step further and say, 'let's resequence the genes' and help them all the way towards a clinical study where the mutation status of a certain gene might be of interest.
Also, for IMGM, the customer really comes first. As we are not operating as a sequencing factory, I think there is a huge benefit in that we are much more flexible, listening to the customer, paying attention to the right size, the right concept for their project. I think that's a very good way to do business. And it's very important for us that even though the technology applied can be very complex, that whatever we hand back to our customer is in a format and form that is understandable, so it's easy for our customer to draw a conclusion from whatever is contained in the data that they generated with us.
What kinds of next-gen sequencing services do you offer, and on what platforms?
Currently, we have four different platforms in house. We have two Roche 454 GS Junior systems and two GS FLX system, and one of our FLX systems has been upgraded to FLX+. We also have the Ion Torrent PGM and the Illumina MiSeq platform.
When it comes to applications, the main focus so far has been on targeted resequencing. We are able to develop complete assays from scratch, so customers can come to us and say, 'we want targeted resequencing done on gene X', and we develop everything necessary, either using amplicon enrichment or in-solution enrichment, for example the Agilent SureSelect system, and then we perform the sequencing.
So far, we have not ventured into ultra high-throughput sequencing, for two main reasons. First, we collaborate intensively with our sister company, the Center for Human Genetics and Laboratory Medicine, and their focus is in diagnostics, so we have been looking mainly towards resequencing applications so far. The other is that for a small company like IMGM, it's quite an effort to go into high-throughput sequencing. I think it will be easier in the future as the bioinformatics footprint gets smaller, the instrumentation gets less costly, and the workflows are more robust and easy to use, which brings down the barrier for a small company to go and do this.
In the past, we collaborated with other companies that supplied short read sequences from ultra high-throughput platforms when it was necessary in the context of a customer project. But we are not currently seen as a player in that arena. I think we are more regarded as a professional, flexible partner when it comes to smaller, specialized projects where it really plays out that we are able to offer everything from start to finish, and where the project management is of much higher value. For example, we try to make sure that the customers have only one project manager who carries them through their projects. So we try to keep everything constant as far as possible, and we make an effort so the customer experience is very good. So far, we have been very successful in that respect; we hardly lose any customers we acquired, so the customer base is stable and growing.
What's your experience with the Ion Torrent PGM and the MiSeq been so far, and where do you see opportunities for these platforms?
We have had the PGM for something like four months and the MiSeq is brand new. We have done initial experiments with the PGM, but we have not really started with the MiSeq.
Both platforms offer a very fast turnaround time, which wasn't so bad with the 454 system anyhow, which has a fairly short run time. Then there is scalability, which is very nice, especially for the PGM. The price is coming down a lot, and we have a high degree of flexibility in what we can put on a platform and how quickly we can change. We will look closely at the data quality and the amount of data but we expect that these platforms make it fairly easy to accommodate small to mid-sized projects, for example clinical or research studies where the targeted resequencing of a number of genes is of importance.
We see, however, that the automated assay design pipelines that Illumina and Life Tech are offering will have their limitations because there is a filtering step used to look for homologies or repeat sequences, and then for whatever you would like to analyze, certain stretches will just drop out, you won't get primer sets to do that, so you miss certain stretches of the sequence of interest. The question is how to fill those gaps. And because we have dealt with assay design from the very beginning, we should be able to elegantly address those holes to offer our customers a complete service, which means that all sequences are actually sequenced.
The PGM and MiSeq are often pitted against each other. Do you see them as competing or do they serve different needs?
They are similar in a way, but I think what is really important is how long it takes from the concept of an application until you actually have it running on the system. That is specifically important for us as a service provider. The initial approach that Life Tech is taking with the AmpliSeq panels and the more or less automated and complete design for a given set of genes is very attractive.
Right now, we have not done test designs ourselves on the MiSeq platform, but our colleagues here in house have done that, and it seems that the initial cost to come up with a design and test it is in favor of the PGM system. Also, if you have to do iterations of the design process, how easy is it to accommodate that, how long will it take, and what will be the price tag of this optimization process? It seems to be, at least from what I know today, that we should be very well off with the PGM system. This would also play in favor of the customers because we could respond to their needs much faster and much more complete.
From what I understand today, the MiSeq will be mainly but not exclusively used for targeted resequencing applications by the molecular genetics department of the [Center for Human Genetics and Laboratory Medicine]. At IMGM, we will still have to see how to implement the MiSeq in our service portfolio. Actually, both companies have common access to all instruments. So it will be a matter of which platform will perform best in which situation or project.
How close are you to offering services on the Ion Torrent?
Right now, we are finishing the runs that we need to become an official certified service provider for the Ion Torrent PGM platform. And then we will also finish the runs for implementing and validating the whole system with a known application, which should give us a very good idea of where we stand with the platform, what the performance and the data quality are, because we have sequenced the same genes with other platforms before. Once that is finished, we will offer Ion Torrent-based services.
Is there any particular application that you would say the 454 platform is suited best for?
We have realized that for metagenomics applications, the 454 platform has a real benefit. But that's also true, for example, for HLA typing, which our colleagues [at the Center for Human Genetics] do in-house here a lot and where the long reads offer a huge benefit. Also, if somebody wants to do amplicon sequencing, looking at exons, but would like to have access to sequence information to the left and right of the exons, it might be a good idea to go onto a long-read platform like the 454.
How many of your employees are bioinformaticians, and what bioinformatics tools do you use?
We currently have one bioinformatician. We are not developing algorithms or specific pipelines. As a service provider, we need to be quick; we don't have the time to play around and try to identify which of the various algorithms for mapping, for example, is the best in this specific case. Much rather, we work in a confined space, we have to get things done as quickly as possible. Therefore, we rely on commercial software packages. Depending on the application, they usually come from the vendors. For example, Agilent has a certain package that we use [for microarray analysis]; and for next-generation sequencing, we use the CLC Bio Workbench. Each of the project managers knows how to use these bioinformatics tools, which have been evaluated and run successfully in those areas. And also, we are a certified and accredited service provider, which means that we work with [standard operating procedures], and therefore, when certain applications are set up, there is a validation phase where the best elements are put together to construct a total workflow.
Do you see microarray-based services being replaced by next-gen sequencing?
It is foreseeable that if the throughput, speed, and price of the sequencing platforms continue to develop as they currently do, there will be a point in time when it's much more beneficial to use a sequencing approach to gene expression analysis and not do it with microarrays anymore.
When this transition will take place is still an open question for me. It may be that for quite a while, microarrays will still be around because everybody knows how to operate them, the instrumentation is out, the cost is not prohibitive, and there are still some things that need to be done on the next-generation sequencing side to make it attractive and surpass microarrays.
Are you planning to add other sequencing services in the future?
I think the time is right to think about other platforms, like, for example, the Ion Torrent Proton, when it comes to whole-exome or whole-genome sequencing. That is something we are currently thinking about, and we are closely looking at how this field is developing and talking to our customer base about how important that would be for them to have as a service.
At the same time, we are developing other applications. For example, we are now getting into metagenomics analysis full-time, which means we are not only offering the sequencing but a complete package, helping the customers to select the right primers to amplify their 16S target sequencing and so on, so it's much easier for them to get the right data back for their research.
Also, originally coming from the context of gene expression analysis, we have a collaboration with Carl Zeiss MicroImaging here in Munich where we explore single-cell or few-cell gene expression analysis [using the Fluidigm BioMark platform]. This might be very interesting for oncology, for example, to look at different portions of a cross-section of a tumor and see how tumor cells might have developed. We are currently in the late phase of our pilot project [with Zeiss] and, presumably, at the end of the year we will offer a gene expression analysis service for single cells or few cells.
A logical extension of that is to do next-generation sequencing on those samples. Next-generation sequencing of single cells is still in an early phase, but that will be something that we will deal with once we have finished the gene expression analysis project.
The vendors seem to believe that many research labs will install their own sequencers as the platforms become smaller and easier to use. How will that affect your service work?
True, if the platforms get more versatile, easier to use, and the cost comes down, that means that many potential customers can do the work themselves. But then again, working with next-generation sequencing is still fairly complex and the platforms evolve so quickly. So even if a customer decides that he wants to buy a specific platform at any given point in time, he still faces the implementation, the training of the people, negotiations with the platform provider about reagents costs, service for the instrument.
Being a service provider, this is our basic background work. We always make sure that the instruments are maintained in a perfect state, so this is something that a customer doesn't have to think about. And also, when it gets more complicated in terms of sequences, we can step in and use our experience, which he may or may not have developed yet. With us, it's a running system, so we can just plug in his project, whereas if he wants to do it himself, he really needs to invest in all those things, and then of the first projects, some will fail. So it takes time, and it's actually months and months until someone says, 'I can do it easily myself.' Therefore, from a cost perspective, for many people it will still make sense to ask a service provider to do it because they get a perfect result, perfect data, at a very reasonable cost, very quickly.
At the same time, I believe that a company like IMGM should develop new workflows and new applications. I mentioned single-cell or few-cell analysis. Metagenomics is maybe not so new but still, if you can do it at a very high level, it's attractive. People are starting to look into the human microbiome, so there are new areas opening up. And then there is the new field of synthetic biology. If you are actually capable of writing code, you have to make sure that what you have written is actually what you want, so I think there will be a lot of sequencing going on in that area. Then there are all kinds of other quality control applications that could be done. Right now, many people, when it comes to next-generation sequencing, mainly think about resequencing human genes, but there may be industrial opportunities, so we are looking into those as well.
To give an example, one way to generate energy is to drill into the ground and get access to hot water, which you can use to generate power. This is a multi-million [euro] endeavor. Unfortunately, there is still life below, and those microorganisms have a tendency to either lead to corrosion of the pipes or to other biofilm-related problems, and it might be interesting, in the early phases of such a plant, to look at the microbial situation and see whether it's OK to bring down regular metal pipes, or whether it needs to be stainless steel or to be reinforced in some other way. It's something that is not a logical extension of how next-generation sequencing has been applied up to now. Thinking about the cost going down even further, maybe new areas of application may spring up where people thought it might be interesting to do this but it's completely out of reach from a cost perspective. I'm pretty sure that will spawn new areas of applying next-generation sequencing.