Can bioinformaticians work from home?

20 years of laboratory journal

Bioinformatics is partly just a collective term under which various sub-disciplines are grouped and not everyone who works with computers in biology or related areas would identify with the term. And here we are not only talking about the computer biologists who deal with biological questions and see bioinformatics more as a method, resource and infrastructure development. Even someone who is 90 percent involved in computer analysis and tends to use service facilities for sequencing sees himself more as a genomicist whose profile includes bioinformatics.

With the boom and the increasing need for data processing, the attitude of life scientists to bioinformaticians also changed. Still ridiculed at the end of the 80s, in the 90s skepticism, envy (symbolizing bioinformaticians as data pirates who live and publish from the work of others) and fear of the new prevailed. The attempts to establish bioinformatics as a service in many places were mostly unsuccessful and so the necessary evil was accepted. At the end of the 90s, stimulated by many funding programs, there was a boom in professorships and in this millennium bioinformatics is seen as an essential part of biological research and is finding its way into almost every laboratory.

If you define the bioinformatician in such a way that he spends more than 50 percent of his time on the computer without manuscript work or communication (e-mail, Skype, etc.), according to internal surveys, more than 40% of all scientists at EMBL are now included - Ascending trend. Bioinformatics is therefore no longer suitable as a category in the citation statistics for the laboratory journal: Too many would complain that they were not taken into account. The upswing in one's own field of science is indeed a satisfaction for someone who was once "smiled at", but it also makes scientific life much more complex and complicated.

In the "golden" 90s, a bioinformatician (at least in sequence analysis) could achieve a biologically relevant result (for example, discover a domain and characterize it functionally) in a few days through a database search and draft an easy-to-structure manuscript; my record was three days from the start of the search to the submission of the manuscript (at the time when the laboratory journal first appeared at EMBL 20 years ago).

This efficiency also had its downsides, because I have already been called a charlatan by members of an appointment committee because, as part of my habilitation, I let a three-week practical course with four students result in a publication: This would not be a real science, where you normally spend years struggling to get the right results and giving students the wrong picture. In general, however, at that time you were a sought-after specialist and could handle pretty much all inquiries (if they were interesting) and because the database sizes were not yet excessive, computing power and storage space were not yet limiting.

Whereas in the past, in a “data-poor” biology, one was mainly concerned with the evaluation of an experimental result, today “big data” is more and more used to generate hypotheses that have to be validated with experiments. Bioinformatics is therefore more clearly involved in the knowledge generation process. The iteration with the experimenters, as is usual in systems biology, is exciting, but results just drag on. The development of the study design with various experimental components implies importance, but also brings more responsibility. Putting a computer experiment in the sand or getting caught up in programming errors may be unpleasant, but it has neither the time nor the financial consequences of a failed complex experiment.

Most of our young bioinformaticians at the institute are in experimental groups and have to take care of their needs themselves, as the respective group leaders have strengths other than computer knowledge. Today many analyzes are web-based on the PC. Often, however, external databases and programs are stored locally (often several times) and the central computers are filled with “jobs” that are not always meaningful and effective.

The work profile of a bioinformatician is broader and goes hand in hand with ongoing training. Just a few years ago you could specialize in an omics method and a corresponding sub-area, and only a few could "script" and analyze. Today both are required and you have to be methodically and biologically fit - also well informed - in order to plug the various existing programs into your own "pipelines". This diversity also means almost daily training and the internet doesn't always help ...

With the first coffee in the office (that was already there 20 years ago), this is also the topic of the first morning meeting: How can you record and bring together 200 bioinformaticians in all corners of the house and provide them with information and avoid redundancy?

We have launched the Bio-IT project, which is supposed to take this into account, but also requires complex logistics: PhD students and postdocs are made aware of bioinformatics in the "routing slip" on the first day of work and receive a corresponding personal message Introduction. Various internal courses, several virtual computer biology centers with service consultation hours in certain focus areas, events to exchange experiences (some with pizza incentives) and an elaborate internal website that has to offer everything from CPU properties, central programs and databases to tips, are part of the BIO- IT package.

Not to forget a working group with members from all scientific disciplines, so that no one is neglected. All of this with the stipulation not to organize too much "top-down" and to assess the needs correctly. When I started there were 20 Hanseln in the institute who saw the computer as the main tool in biology, and we played table tennis after lunch. That too was “community building” and clearly “bottom-up”; the redundancy was minimal like the data we handled, but that is no longer possible from a certain institute size onwards.

The message from the cell phone (automatically initiated by the institute-wide calendar system) reminds you that a telephone interview with a journalist is to be arranged next (public relations is becoming more and more important and the relevant department already has more than ten people - 20 years ago there was no one there) . It's actually nice when people are interested in our work and the term bioinformatician can also be admired on television. But since there is an appointment of this kind almost every week, that adds up to a lot of time that is missing from the actual work, bioinformatics ...

After dozens of signatures, smaller decisions and a few short one-on-one discussions, the highlight of the day in the late morning: a real scientific discussion, a real “brainstorming” session about a project that requires creativity. Because of the increased complexity of the data, it is easier to work in small teams in which the members have different training profiles. Putting your heads together is usually fruitful. However, here, too, it is mostly about the clever “cleaning” of raw data, which must be made comparable, about the optimization of programs that enable an analysis in a reasonable time, and about the independent confirmation of results. Since nobody knows the truth, bugs are hard to find and good tests are essential.

An important point is the interpretation of the results every time. In a comparative analysis one is happy about correlations, but these are often only indirect. It has already happened that the observed change in the intestinal flora in patients has nothing to do with the actual disease they are targeting, but with a drug that many patients have to take, or with changed eating and lifestyle habits after one Diagnosis. But at least it is fun to tinker with scientific questions, you have been trained for this for a long time - nobody has ever taught me how to keep enough time for science when organizational requirements increase ...

After noon, the organizational activities in the next block of e-mails say hello immediately: "Time sheet" reminder from the HR department, need for clarification in the case of a computer purchase by the finance department, an inquiry list from the local ethics committee regarding an experiment, a request from the "Grant Office" because of allocated funds, and our legal department has written a 12-page document to clearly regulate cooperation with a university abroad.

Our Technologietransfer GmbH does not fully agree with this, as some clauses look like service and therefore have to be dealt with through them - according to internal rule number 1342 [number changed by the editorial team].

Sure, I earn more than my postdocs and thus I also earn this type of e-mail, but postdocs also get some of the increased professionalism these days: they have to write many internal reports, and are allowed to drive to arranged meetings within the framework of research grants Documenting your working hours electronically by the hour (the EU administration sends its regards), that certainly does not count towards bioinformatics and does not fit into the cliché of scientific freedoms. 20 years ago I had to send in my report after finishing my research grant and I was never sure if anyone would ever read it, but that was after two or three years of undisturbed research ...

Despite all the professionalism, my scientific experience is in demand again in the early afternoon: A publication is being prepared and priorities have to be set and illustrations have to be improved. I can complain about that now. For example, that most biologists are still at war with box plots and that one has to simplify things for the reader. At least that's my impression, not much has changed here, apart from the fact that software simplifies the creation of images - if you can use it. Back then, as a postdoc, I also needed advice on the presentation and received it ...

Now it's time to disappear home unnoticed, because I have to spend two hours quietly editing a manuscript on genome annotation and tinkering with a grant for diagnosing stomach cancer. The e-mail block in the late afternoon speaks a different language: three times "Dear Sir", that is, postdoc inquiries from India, twice a request for a review for a journal and once for a research grant. In addition, several invitations to commercially organized conferences and offers from all sorts of companies to help with project or grant management or with bioinformatics yourself.

So it is not only in the institutes themselves that more science is organized (with the same number of scientists). There is also a growing economy flourishing around science. 20 years ago there were only the scientific publishers, whose expensive specialist journals were very profitable, among other things, due to mutual peer review without payment. Newcomers like the free laboratory journal were very welcome back then ...

In the evening it goes to the airport, the travel activity has not changed that much, apart from the fact that you now put your lecture together on the plane. The fact that there will be more is due to the job profile, as a postdoc I made ten trips a year, but 20 years ago as a group leader I already had around 20, but almost all of them went to scientific conferences. Today, the CO2 emissions by air are not much higher, but a trip is often linked to several events. Conferences are increasingly being replaced by appraisals, project or management meetings.

The question arises, where is the journey going ... with bioinformatics. Will biology be as quantitative as physics? Will the bioinformatician job profile still exist, since every biologist has to have bioinformatics to hand as a craft, like word processing or presentation programs today? Or will bioinformaticians be more in demand than ever because data generation is becoming cheaper, but analysis is becoming more and more demanding? Both of these suggest that acquiring bioinformatics skills is a good investment. But what will bioinformatics look like in everyday life? Is all experimental data managed, networked and bioinformatically analyzed anonymously in the “cloud”? Will a better understanding of data together with “big data” lead to a culture in which failures are also published in order to get better statistics (there is still a huge problem in biology and other natural sciences)?

Actually, bioinformatics is much more exciting, more diverse - also more successful - than it was 20 years ago, and it can be used across disciplines. Compared to the "good old days", a bioinformatician now achieves many more and more important results per unit of time, despite all the "professionalism" that hampers creativity and efficiency in everyday work.

It would be great if I could start again today as a doctoral student ...

Peer Bork is head of the Structural and Computational Biology Unit at the European Molecular Biology Laboratory (EMBL) in Heidelberg.


Last changes: 07/11/2014