BACK TO THE FEATURE INDEX

To follow up on last week's feature on bioinformatics, Next Wave thought you'd like to know where the field is headed and what you're going to need to do if you want to stay ahead of the curve. So, we asked the bioinformaticians we interviewed for the feature to gaze into their crystal balls and make some predictions. Here--lightly edited--is what they said.


My vision for the next 5 or 10 years encompasses not only an increasing flood of new data to be processed on computers of ever-increasing speed and storage space. No, I think we will return with all this information to the domains of classical and modern biology ... physiology; biochemistry; genetics; evolutionary biology.

Bioinformatics itself, I think, will become part of theoretical biology, which until now has not provided a conceptual framework for the discipline. Things are different, say, in physics, where the doctrine of modern physics is written in the language of its theoretical foundation. Theoretical biology, with biomathematics and biostatistics, has been--at best--an ancillary discipline, an illustrative topic. But the recent development of genomics and genetics could transform theoretical biology into the conceptual basis for all biology. Like theoretical physics, modern theoretical biology can explain and integrate, associate and generalize. Nevertheless, empirical work--observations of, or experiments on, nature--must remain the ultimate source of biological knowledge.

--Jens Reich (Group leader of bioinformatics at the Max Delbrück Center and professor of bioinformatics at Humboldt University in Berlin) Text derived from Reich's article in the 1 September release of Next Wave.


In terms of who will be doing bioinformatics in the future, the answer is "You all will be," so you'd better get in gear. Anyone doing any molecular biology certainly needs to be doing bioinformatics--it should be part of your skills set.

What are the new things that are happening? The human genome is coming out right now, and that is a resource that we'll only understand by using bioinformatics. But the thing to make quite clear is that we're not actually finished with the human genome and we probably never will be. That's partly because the sequence is still being refined, but it is also because as we obtain other mammalian genome sequence data we're beginning to carry out comparisons. We know that genes need to be conserved or nearly conserved, so we'll find lots of things we expect. But in the so-called junk DNA--98% or so of the genome--it's likely that we'll find other conserved regions that aren't in genes. We would assume that any such conserved regions are biologically important; otherwise they would have diverged over time. Figuring out what they do will be an interesting challenge, and it will take some time.

--Graham Cameron (Joint head of the European Bioinformatics Institute)


What would I do if I were starting out in bioinformatics today? Well, let me restate the question--What are the bioinformatics problems we are facing today, and what will be the next set of challenges?

The soon-to-be-released complete human genome sequence is just one set of data. The mouse sequence is nipping at the human's heels, and the rat genome is skulking under a box somewhere nearby. But associating gene information with related genomic results is a constant challenge, especially when the data are appearing faster than we can manage them. Scientists trained to manipulate these large data sets, build data relationships, and classify and organize the information are already in great demand, and they will be for quite a while yet, I think.

At the nucleic acid level, understanding the precise regulation of genes through analysis of gene expression data will be of utmost importance. But the primary sequences of the human, mouse, and rat genomes are also tools from which we can derive the protein sequence for each gene. So, my guess is that there will be an explosion in the computational prediction of three-dimensional protein structures, as well as in the analysis of empirically determined protein structures. If we can understand how the primary sequence and the secondary and tertiary structures map to function, we will be closer to understanding how proteins work. With this understanding, we'll be better able to predict how variations in a protein might lead to changes in phenotype, which--when we're talking about variations associated with dysfunction and disease--will be crucial for figuring out how drug treatments might differ among individuals. What I predict, then, is that the bioinformatics revolution will help to establish personalized medicine--getting the right drug to the right person at the right time.

As for skills and qualifications, a comprehensive statistical background will be needed for this kind of research. So, I'd better sign up for a refresher course in biostatistics as soon as possible!

--Martin Leach (Director of bioinformatics, Curagen Corp.)


Bioinformatics covers huge areas of expertise these days, so I'll focus on my own particular area: using protein structural predictions to link DNA and protein sequences to their functions.

The number of experimentally determined protein structures is quite small at the moment, but this will change markedly over the next 5 years. The flood of newly solved structures will revolutionize some areas of research that are currently very difficult to work in due to a distinct lack of data. At the top of my list has to be the prediction of interaction surfaces, because it is through these surfaces that all proteins achieve their functionality.

As for acquiring skills, my suggestion is that anyone wanting to enter bioinformatics from the biology side needs to become a proficient computer programmer. This may not be as arduous as it sounds, because more often than not, the same basic algorithms are reused in new applications. So you can learn a lot just by studying and understanding existing algorithms.

--Mark Swindells (Scientific director, Inpharmatica)

I believe that bioinformatics will be of growing importance. Already, biosciences eat up most of the computing power in the world. There's work in bioinformatics till I'm 150!

--Arne Stabenau (Scientific programmer at the European Bioinformatics Institute)