Skip to main content
Premium Trial:

Request an Annual Quote

Let the Bioinformaticists Save Themselves


Les Klimczak is the director of bioinformatics and information technologies at Psychiatric Genomics in Gaithersburg, Md. His research interests include data mining/knowledge discovery in databases, integration of biological and chemical data, and applications of genomic information to drug discovery.


It is hard to be unsympathetic to the recent predicaments of bioinformatics tool and service providers. However, I think that it is unfair to generalize their problems to the entire bioinformatics community and to reach the alarmist conclusion that bioinformaticists need “to be saved,” as was the opinion expressed on this page in July by Martin Gollery. Lest bioinformaticists be relegated to the status acquired recently by Enron executives and Andersen accountants, we must affirm a more constructive and positive view of the current role of bioinformatics in industry. But this may also be a good opportunity to discuss the expectations for bioinformatics in genomics, biotechnology, and drug discovery and to help chart a future course that would avoid the mistakes of the past.

Bioinformatics is extremely successful where it is used to support directly the mission of research — as an integrated component of accelerated discovery by generating and analyzing large biological and chemical datasets. Many bioinformatics-related companies that have strong products that facilitate this mission are doing well and enjoy the respect of their customers: Partek, GeneSpring, SciTegic, and Spotfire are just a few examples.

The problems start when the marketed bioinformatics solutions become an end in themselves instead of serving as the means to reach discovery goals. A successful internal bioinformatics discovery program must rapidly deliver actionable results, such as gene or drug candidates, to downstream customers to move the company forward. The problem-specific discovery processes require customized solutions that can only be delivered by problem-oriented development and integration of best-of-breed tools, which will move the problem directly from point A to point B. Maintaining a pharaonic enterprise system that provides only very generic functionalities will not contribute to these goals but will divert resources from the analytical work; the goal-oriented analytics will also be much more difficult within the constraints of such a system. This will particularly affect startups and biotechs, but not even big pharma companies, as their budgets tighten, can afford to follow a circuitous route to discovery.

To be considered a part of the solution instead of the problem, a good bioinformatics product must have a well-defined functionality and provide significant added value. Programmable advanced analytical modules that can be plugged into scalable and extensible pipelines are winning over inflexible GUI wrappers around public tools. We still need improvement in how the analytical tools interface with each other and the data stores: LIMS products should be more flexible in modeling real-life objects and exposing them to programmatic access.

The wounds of the unsuccessful bioinformatics companies were mostly self-inflicted: their monumental and overpriced solutions did not help to accelerate the discovery process and they lacked long-term viability. Marketing hype and sales arrogance did not help — it was a red flag drawing attention to technological weaknesses. It is ironic that those of our colleagues in discovery informatics who had the least political clout to resist the relentless sales pitches channeled through non-technical business management were the first to be blamed for the lack of deliverables and often had to re-enter the job market side by side with the developers of those half-baked solutions.

In the long term, commercial bioinformatics will gain from the elimination of bad ideas by the marketplace as it will be able to better focus on the ideas that work. As with any commercial ventures, bioinformatics ideas must be fit to be able to survive in the marketplace.


Opposite Strand is a forum for readers to express opinions and ideas about trends and issues in genomics. Submissions should be kept to 550 words and may be submitted to [email protected]

The Scan

Harvard Team Report One-Time Base Editing Treatment for Motor Neuron Disease in Mice

A base-editing approach restored SMN levels and improved motor function in a mouse model of spinal muscular atrophy, a new Science paper reports.

International Team Examines History of North American Horses

Genetic and other analyses presented in Science find that horses spread to the northern Rockies and Great Plains by the first half of the 17th century.

New Study Examines Genetic Dominance Within UK Biobank

Researchers analyze instances of genetic dominance within UK Biobank data, as they report in Science.

Cell Signaling Pathway Identified as Metastasis Suppressor

A new study in Nature homes in on the STING pathway as a suppressor of metastasis in a mouse model of lung cancer.