I think you must distinguigh two types of biologists (I'll describe the extremes; there are all the shades of gray inbetween). One type studied biology just because it had something to do with flowers and bees and is all so natural and ecologic. Such biologists had a focus on biology in school because they anyway never understood maths, and chemistry and physics were too tough. The other type is interested in "how nature works". Biology for them is the top-level expression of physics and chemisty, leading to the most stunning properties, and math is a useful tool to explore, formulate and describe models of the living world. These scientists do not have any problems with maths, formulas, programming and so on. They just use it when they need it.

To my opinion, "making bioinformatics more interesting" is pretty much worthless for the first kind, and it is not required for the second kind. We more and more dive into a habit of "selling" and "promoting" our fields of science, as if it was a product. I think this is not the best way. Scientists are interested in their topics, they want to learn something. This includes searching for smart methods that may help. If such a method was math, statistics, bioinformatics, or chemistry, quantum physics or whatever - the scientist will anyway try to learn this and/or try to collaborate with experts of the respective fields.

Further, biologists do rarely need to understand how bioinformatic tools work. It is just important that they do its jobs and that the biologist can correctly interpret the results. The is no need to understand the structure and organisation of a data base to search for information, no need to understand BLAST algorithms if all that is wanted is a ranking of similar sequences.

I think this post was provocative enough to get some interesting reactions :)

I think you must distinguigh two types of biologists (I'll describe the extremes; there are all the shades of gray inbetween). One type studied biology just because it had something to do with flowers and bees and is all so natural and ecologic. Such biologists had a focus on biology in school because they anyway never understood maths, and chemistry and physics were too tough. The other type is interested in "how nature works". Biology for them is the top-level expression of physics and chemisty, leading to the most stunning properties, and math is a useful tool to explore, formulate and describe models of the living world. These scientists do not have any problems with maths, formulas, programming and so on. They just use it when they need it.

To my opinion, "making bioinformatics more interesting" is pretty much worthless for the first kind, and it is not required for the second kind. We more and more dive into a habit of "selling" and "promoting" our fields of science, as if it was a product. I think this is not the best way. Scientists are interested in their topics, they want to learn something. This includes searching for smart methods that may help. If such a method was math, statistics, bioinformatics, or chemistry, quantum physics or whatever - the scientist will anyway try to learn this and/or try to collaborate with experts of the respective fields.

Further, biologists do rarely need to understand how bioinformatic tools work. It is just important that they do its jobs and that the biologist can correctly interpret the results. The is no need to understand the structure and organisation of a data base to search for information, no need to understand BLAST algorithms if all that is wanted is a ranking of similar sequences.

I think this post was provocative enough to get some interesting reactions :)

Thank you Dr Jochen Wilhelm for your reply. I think i choose the wrong title of my question, esp the word "BEST". Yes, i completely agree with you that there is no need to get the BEST of anything / any branch of science. And yes, every field has its own interesting impacts depending upon the requirement. your point of 'selling/promoting' of fields is really thought provoking and stress on the best of doing science. But here i also want to say that "my motive was not to make bioinformatics more IMPORTANT than other branches of science", My focus was "Since a lots of experimental biologists wants some tools to get some help in interpretation of their studies /to analyze / to corroborate the results in more conclusive way etc" ....... i want to know those problems which they think that it can have some intelligent guess by using computational method/bioinformatics (regardless of whether they understand that tool or not)..........M sorry if my question directs towards any wrong way of doing science.

Jochen, you make some excellent points. While I agree that neither kinds of biologists need to go through the equations, there is definitely a widespread lack of appreciation at a higher qualitative level, such as understanding comparative evaluations of methods and knowing/caring what tools to apply, how and when.

BLAST is actually a great example to have picked Ravi, as it is widely abused and the E-value is one of the most widely misinterpreted measures in biology, certainly in sequence analysis. Of the second type of biologist that you mention Jochen, few could tell you what an E-value means, despite almost all of them being capable of understanding and no doubt most of them having used them to do science that they have published.

BLAST for amino acid sequence analysis is now pretty much defunct, since there are far more sensitive methods (such as JACKHMMER and HHsuite tools) that are now just as fast to run. The momentum that the community has in using BLAST, is damaging to science and progress, and as Ravi suggests in his post, this is indicative of a wider issue that must be addressed in the biological community: somehow the interest level in bioinformatics tools and their application must be raised, because scientists really care about the value of their output, but for reasons that I will not try to get into here, work is routinely published with a bioinformatics component or dependency that is false, wrongly applied, or just deplorably far from the state of the art.

As for what can be done about it? I suppose referees and journals need to raise standards, particularly in good biology papers with a poor bioinformatics component. I am guilty myself of letting through papers where the bioinformatics component is scientifically poor or even irrelevant, where I have judged it not to have significantly impacted their final biological results (although the authors clearly had no idea what they were doing).

Thank you for such a wonderful discussion. I agree as biologists, we need to expand our knowledge in the field of bioinformatics. I hope that both biologists, medical researchers and bioinformaticians, computer scientists will work side by side to publish the bio-medical researches as the bioinformatic component is obligatory for the high throughput data analyses derived using present day hi-throughput technologies such as microarrays. Presently, I have limited resources, so my medical research is at a pause, but as a medical researcher, very early on I have realized the need for all these communities to come and work together closely as a team. The referees and 'journal suggestion' is a good one too from Julian Gough.

Statistics was born inside biology as a matter of fact ANOVA still maintains in its structure the frame of experimental fields where two (or more) factors are varied continuously along experimental parcels and we call Y the dependent variable because is the initial letter of Yield. Until the 50's and 60s any physiologist will know what a Fourier spectrum was and in very minor and neglected biological fields like phytosociology and plant ecology the scientists use very refined techniques like multidimensional scaling and ordering methods that many physicists and engineers consider as stellar math (try and ask a medium applied mathematician how to generate an explicit space from between elements pairwise distances and compare their answers with this: http://ordination.okstate.edu/).

This kind of quantitative approaches, even if in many case use very similar techniques than the math used in physics are totally different in their basic phylosophy: the computation start AFTER the data are collected, while in physics the mathematics comes BEFORE, in this respect statistics (and bioinformatics) in my opinion are not theoretical biology in the same way than theoretical physics, they are 'tools for thinking' in an artisan way. This was almost totally lost when biology became an affair of looking at spots variously scattered on a photographic paper (Northern, Western..blots) and making annoyng lists of if X is involved in Y that in turn is involved in Z..., on the other side mathematically inclined persons when looking at biology completely loose of relevance (are more than 70 years than biomathematical journal reiterate Volterra and Lotka equations) because they want soon or later arive to write an equation instead of using math as a microscope.

Remedies ? Stopping the ultraspecialization of sciences and coming back to teach the essentials in a single basic CV in science common to all and specializing only when having acquired the common bases.....

Jochen Wilhelm view are really aspiring.

All comments are equally fruitful ...

Jochen's view is indeed provocative! My feeling on this matter is that bioinformatics is a young discipline, which has evolved and is still evolving as computering science and power proceed. But don't forget that bioinformatics was used and developed very early by population geneticists because this field needs math equations, statistics and models. Many of them like flowers, bees and also maths! On the other hand, molecular biologists had no real need for bioinformatics until the mid 90's. They had enough with cloning and sequencing genes and mRNAs, making cell lines grow in culture, development of antibodies and other techniques. No need to perform statistics, except triplicate comparison. Bioinformatics was just a tool merely used to find restriction sites in DNA sequences or some phosphorylation sites in proteins...

The current situation is completely different. It's now so easy to generate millions of data (DNA, RNA, protein sequences, images, gene frequencies, SNPs ...) that every one needs bio-informatics to extract relevant information (e.g., as Anupam said for NGS). One might still want a ready-to-use tool but I feel that no one should use a tool without an idea on what it does and how setting parameters influence the output. And this is what's happening: experimental biologists have now a better understanding in optics, image acquisition and processing, statistics, RNA-Seq analysis because they need it.

But it's true biologists don't know how to program and this is where they can meet with bio-informaticians, as they meet with physicists when they collaborate to understand how the physical properties of biological matter contributes to cell properties. Meanwhile, bioinformatics has now its own field of research and should not be considered any more as a tool provider.

So things are quickly moving, all the more that younger scientists are more than comfortable with computers. As John said, research in biology won't offer the same lanscape in a few years.