The changing marketplace of bioinformatics

 

Nature Biotechnology  18, 1247 - 1249 (2000)
doi:10.1038/82351

The changing marketplace of bioinformatics

Paolo P. Saviotti, Marie-Angèle deLooze, Sylvie Michelland & David Catherine

Paolo P. Saviotti is research director (saviotti@grenoble.inra.fr ) Marie Angèle deLooze is research engineer (delooze@grenoble.inra.fr ), Sylvie Michelland is post-doctoral researcher and David Catherine is a post-graduate student at INRA-SERD, Université Pierre Mendès-France, BP 47, 38040 Grenoble Cedex 9, France .

The demand for more sophisticated bioinformatics solutions is expanding opportunities for biotechnology startups, but traditional IT companies are also moving into the market.

 

Bioinformatics aims to turn the massive volumes of data arising from industrial-scale biology and chemistry into usable information for researchers in basic and applied research. Traditionally, the specialist market for bioinformatics has been occupied by a small number of companies providing expensive software to single clients that addresses isolated problems in the drug discovery process. Analysis carried out by our group, however, suggests that the bioinformatics marketplace may be evolving, with companies that focus on information technology (IT) in other industries moving into the arena.

Emerging niches
Several roles are evolving for bioinformatics in the life sciences. First, with the drive toward high-throughput procedures in the laboratory, it is playing an integral part in the automation of biological and chemical research. Second, it is providing tools for the capture, management, and analysis of disparate data rapidly accumulating from these industrial-scale approaches. And third, it is facilitating the exchange and dissemination of information between isolated groups in large organizations. For example, Genentech ( S. San Francisco , CA ), Bayer ( Leverkusen , Germany )¹, and Pharmacia Upjohn ( Kalamazoo , MI )² have all recently reorganized their R&D activities to address problems with knowledge management and the poor information exchange between drug discovery and development teams.

Several niches appear to be emerging in the bioinformatics market. First, companies, such as Spotfire ( Cambridge , MA ), MDL ( San Leandro , CA ), and Silicon Genetics ( San Carlos , CA ), are providing software packages for data visualization, interpretation, and analysis. Although this market is well established, it is relatively small, there are only a limited number of customers, the clients' needs are diverse, and academics and other researchers are commonly eroding markets by distributing competing software freely.

A second group of companies, which includes DoubleTwist (Oakland, CA), e-Bioinformatics (Pleasanton, CA), and Compugen/LabonWeb (Tel Aviv, Israel), are offering comprehensive bioinformatics platforms accessible to life scientists over the Internet. With an estimated market of around 150,000 life scientists as potential subscribers, these companies (often termed application service providers) are also hoping to supplement revenue through e-commerce, with links to equipment and reagent suppliers such as Chemdex and SciQuest.

A third model is to rent out data architectures/infrastructures and bioinformatics expertise to large pharmaceutical clients that are struggling to integrate disparate databases and technology platforms. These companies include not only dedicated bioinformatics companies, such as Lion BioScience (Heidelberg, Germany), who closed a $50−100 million landmark deal with Bayer in June 1999, but also more conventional IT companies such as IBM and Motorola (see "Bioinformatics patents" below).

Research-based companies are also offering bioinformatics tools along with their content; companies that fall in this category include Incyte Genomics ( Palo Alto , CA ), Celera Genomics ( Rockville , MD ), and Curagen ( New Haven , CT ). These companies either license their data and bioinformatics tools to drug companies or offer access to academics/companies on the basis of subscriptions.

What is a bioinformatics company?
Bioinformatics is an emerging specialty. At the moment, it cannot be considered either a discipline or an industrial sector. Most bioinformatics firms are small biotechnology firms that initially developed IT tools to deal with the enormous amount of data generated by internal biological R&D and subseqently capitalized on that investment by selling the tools and related services to outside clients. The number of applications of IT to the life sciences is very large and it is clear that only a fraction of the potential has been realized so far. Thus, our definition of a bioinformatics firm might be very different 10 years from now. Furthermore, many firms that presently have bioinformatics capabilities solely for their internal research might also rapidly transform into bioinformatics firms.

Our database, which was built from data extracted from identity cards of firms and other descriptions, contains 135 firms, as of November 2000. Of these, the vast majority (84) are based in the United States . Forty-four of the remaining firms are based in Europe, four in Canada , and one in the Middle East . We believe that this bias reflects the general lag in European startup activity behind the United States for a new technology area. According to Ernst & Young³, European firms account for 40% of general biotechnology; our survey indicates that only 32% of bioinformatics companies are based in Europe , reflecting the nascent state of the field. We can expect the geographical distribution of firms to become less skewed and show greater European representation as the technology matures.

This interpretation is reinforced by the relative youth of most bioinformatics firms. Of the 129 firms in our sample for which the date of foundation is known, most were created after 1992. The annual rate of creation for bioinformatics firms increases after 1992 to an average of 14 new firms per year, with a maximum of 25 per year in 1997. Before 1992, only 2 firms per year were created. With 23 new companies in 1998, there has clearly been an explosive growth in bioinformatics startups.

Our data also suggest that most bioinformatics firms are small- or medium-sized enterprises. Only the four largest firms in our sample have more than 500 employees, 25% of them have between 100 and 500 employees, and 60% have between 10 and 50 employees. Of the existing firms, 61% specialize in bioinformatics and 39% include bioinformatics among their products.

Nevertheless, for the past two years, we have observed an increase in specialized firm creation (70% of specialized firms against only 30% of nonspecialized). This growing predominance of specialized companies is expected as to continue as the size of the market for bioinformatics continues to grow. The future stability of this pattern of specialization may depend on the supply of bioinformaticians. This means that their knowhow is in great demand and that it is very difficult for pharmaceutical companies to recruit bioinformaticians for in-house efforts. This may have implications for bioinformatics companies, as the shortage of expertise may force drug companies to outsource their bioinformatics needs.

We consider our sample very large with respect to the population of bioinformatics firms at the time of the study, but not necessarily complete. Many biotechnology companies are intensive users of bioinformatics and could easily become producers. Thus, the population of bioinformatics firms could change very rapidly. On the other hand, an inspection of the patent literature reveals many "nontraditional" companies are moving into the bioinformatics area (see below "Bioinformatics patents").

The market
Only a few estimates of the bioinformatics market exist. Frost & Sullivan estimate that worldwide bioinformatics sales should reach $160 million this year, with the potential to grow to as much as $2.0−5.0 billion in the next five years.

We have evaluated the market size by means of the turnover of existing companies. By necessity, we assumed a constant turnover/employee ratio and calculated an approximate turnover for all the firms in our sample. The total market size calculated on this basis is $917 million for the end of 1997. Considering that Ernst & Young estimated that the total market size for all biotechnologies is $20 billion, our estimate suggests that bioinformatics could account for about 5% of all biotechnology revenues.

These numbers probably underestimate the use of bioinformatics, because certain pharmaceutical and biotechnology firms have extensive in-house bioinformatics expertise, which never offers its services for sale.

Customer base
One of the recurring problems in bioinformatics is that it has been difficult to find a broad audience for a sufficient number of software applications to support a business. The types of customers who use bioinformatics products not only include traditional pharmaceutical and agricultural researchers, but also increasingly academics who are carrying out in silico research.

According to our database, bioinformatics customers are distributed among several industrial sectors and applications. The pharmaceutical industry, the human genome project, and biopharmaceutical industry together account for over 40% of all bioinformatics sales. The next largest technology sector is agricultural biotechnology, accounting for nearly 11% of overall sales. The main bioinformatics markets are shown in Figure 1. As the genomics projects progress, and more and more basic and applied researchers shift focus from wet biology to in silico experiments, it will be interesting to follow the importance of bioinformatics in different technology sectors.

 

Bioinformatics patents
To further assess commercial activity in bioinformatics, we searched the Derwent Biotechnology Abstracts (DBA) database for patents containing the words "computer, computing, DNA chip, biochip, gene chip, bioinformatics, or informatics." An inspection of the data reveals that the number of bioinformatics-related patents has been increasing steadily from 1979 to 1997, after which there has been a noticeable boom in patent applications, with a peak (60 patents) in 1998. When the data are tabulated (see Table 1), three main types of patentee are discernible: First, companies and institutions interested in applying IT to production processes, such as the control of fermentation reactors; second, pharmaceutical companies, bioinformatics startups, or public research institutes interested in applying IT to R&D processes (e.g., sequencing, biochips, drug design, and mathematical modeling); and third, companies and instrumentation firms interested in creating the tools and solutions for R&D processes.

 

 

The first group of patentees not only file much older patents (1983−1992) than those of the second and third groups, but also apply IT to the automation of a production process. These institutions are commonly either Russian or Japanese. Hitatchi is among this group, although it recently has also been issued bioinformatics patents related to biochips.

Companies involved in the second type of patent application tend to be biotechnology companies with a heavy investment in research (e.g., Affymetrix, Affymax, or Human Genome Sciences). They are often involved in genomics and are biotechnology startups (and to a less extent pharmaceutical companies) that have developed bioinformatics competencies in tandem with their research programs. The date of patents issued to these companies ranges between 1994 and 2000. These patents are thus much more recent, corresponding to the young age of the filing companies.

But perhaps the most striking finding is the appearance as bioinformatics players of firms such as Motorola ( Chicago , IL ) and Kodak ( Rochester , NY ), which have traditionally been involved in mainstream electronics, IT, and telecommunications. Similar to the second group, patents issued to these companies range from 1992 to 2000, which is unsurprising because the two groups of patents deal with very similar or complementary knowledge fields and applications. If IT-based companies continue to invest in bioinformatics, we can expect the average age of patents to fall rapidly.

Recently, other mainstream IT companies, Hewlett Packard (Palo Alto, CA), IBM (Armonk, NY), COMPAQ (Houston,TE), and Hitachi (Tokyo) have also invested in bioinformatics. Both Agilent Technologies, a subsidiary of Hewlett Packard, and IBM created their Life Science Business Units^{6, 7}. In addition to internal investment large IT companies are involved in collaborations with specialist bioinformatics firms and start-ups. IBM is now marketing DiscoveryLink, a software platform for linking sets of disparate kinds of data into a single computing environment for use as a potential standardized system in collaboration with the bioinformatics startup NetGenics ( Cleveland , OH ). Also, IBM signed an alliance with Incyte Genomics for the use of Discovery Link by the latter.

Not all IT based firms have the same entry strategy⁷. The firms mentioned above invest internal resources in R&D in bioinformatics and form alliances with bioinformatics firms. For other IT based firms, such as SUN and Silicon Graphics, bioinformatics is simply an emergent and rapidly growing market demanding extremely powerful super computers.

The movement of these IT-based entrants into the market is important because they are very large and powerful firms capable of shaking up the industrial structure of bioinformatics. It is tempting to speculate that the expertise of these companies in other industries might be rapidly translated to software solutions that provide the kind of standardization, integration, and analysis of the data so sorely needed. That said, software will only be standardized when the underlying cognitive tasks have become clearly established and demand only quantitative extensions of existing concepts. At this stage, it is unlikely that IT-based companies will be able to integrate all the biological competencies required to become the only actors involved in the production of bioinformatics. However, we can certainly expect that they will exert an important influence on the future corporate structure in bioinformatics.

Conclusions
In general, bioinformatics companies are currently small or medium in size, very young, and have entered the market with knowledge of the biotechnology arena. With the recent entry of firms such as Hewlett Packard and IBM, and the IP portfolio in bioinformatics acquired by Motorola and and Kodak, the dominance of biotechnology-based startups may be coming to an end, especially given the power of the IT-based entrants. While this is not necessarily going to eliminate biotechnology based start-ups, it is likely to recreate from the IT side a pattern of alliances similar to those existing between pharmaceutical firms and biotechnology start-ups in the life sciences.