|
From
•>>April 2005
Bjarne Knudsen answers
a few questions about this month's emerging research front
in
field of Computer Science: Computer Science
Article: RNA secondary structure prediction using stochastic context-free grammars and
evolutionary history
Authors: Knudsen,
B;Hein, J
Journal: BIOINFORMATICS, 15: (6) 446-454, JUN 1999
Addresses: Aarhus Univ, Inst Biol Sci, Dept Ecol & Genet, Bldg 550, Ny Munkegade, DK-8000 Aarhus C, Denmark.
Aarhus Univ, Inst Biol Sci, Dept Ecol & Genet, DK-8000 Aarhus C, Denmark.
|
|

Why do you think your paper is
highly cited?
|

“One of the most important aspects of this new RNA structure prediction method is that it utilizes an explicit model of sequence evolution.”
|
|
The paper deals with a new RNA secondary structure
prediction method based on evolutionary analyses, which is in
the interesting boundary region between computer science and
biology. Classical computer science methods are used to help
solve a problem of interest for molecular biologists. This
results in a number of citations from both algorithmical and
more applied papers. Another reason for our paper being highly
cited is that it provides a web-based method for RNA analysis,
which is easy to use, and thus fairly widespread.
Does it describe a new discovery or new methodology
that's useful to others?
The method described has helped a number of scientists
infer secondary structures for RNA molecules, leading to a new
functional understanding. The method has, for example, aided
in the determination of the RNA structure of the leader
sequence in the HIV genome. The predicted structure has been
experimentally verified by the research group of Jørgen Kjems
at the University of Aarhus, Denmark. Present experimental
work is focusing on the biological importance of these
structures.
One of the most important aspects of this new RNA structure
prediction method is that it utilizes an explicit model of
sequence evolution. By applying this method, it becomes
possible to make much more reliable predictions, due to the
additional information which can be extracted from related
sequences. Including evolutionary information in
bioinformatical analyses is a promising approach, often
leading to very good results. With the numerous genome
projects being undertaken, this additional information is
becoming easily available, and good algorithms for its usage
are essential.
How did you become involved in this research?
Jotun Hein and I started developing the RNA structure
prediction algorithms at the University of Aarhus in 1998,
resulting in a theoretically oriented paper in the journal Bioinformatics
in 1999 (15(6):446-454). The present paper improves and
further evaluates the method. Our work was inspired by work on
protein structure prediction done by Jeffrey Thorne, Nick
Goldman, and David Jones in 1996.
Working in the research group of Jotun Hein was very
inspiring, both due to his excellent scientific abilities and
also because of the opportunity to work with the group he had
assembled—a team which included biologists, molecular
biologists, computer scientists, and statisticians. This
provided a perfect combination of researchers for
bioinformatics research since the field is quite diverse. At
present, Jotun is a professor at the University of Oxford. The
RNA paper was published while I was working at the University
of Florida with Michael M. Miyamoto, where we continued to
develop a wide range of bioinformatics algorithms.
At present, I am the chief scientific officer in a
bioinformatics software and consulting company called CLC bio
( www.clcbio.com).
We aim to further develop the RNA structure-prediction
algorithms, along with other algorithms for DNA, RNA, and
protein analysis. One of our main interests and challenges is
to produce user-friendly software, making the advanced
algorithms available to general molecular biologists and
biochemists
Could you summarize the significance of your paper in
layman's terms?
Chemically, RNA molecules closely resemble DNA, but their
biological roles differ significantly. DNA is exclusively used
to store information, while most biological functions are
undertaken by proteins. RNA has an intermediate role, both
participating in the storage and transfer of information and
in more directly functional roles.
RNA is involved in a number of the most fundamental
cellular processes, including the whole machinery involved in
reading the information from DNA sequences and building the
encoded proteins. This central role of RNA has led researchers
to suggest that RNA may have formed the basis of early life on
Earth, before DNA and proteins were used by biological
systems.
The function of an RNA molecule is tightly linked to its
structure. The biological importance of RNA has of course led
many scientists to study the structure and function in great
detail. It is, however, a very time-consuming process to
experimentally determine the structure. Thus, computational
methods, including the one in our paper, have proven very
useful in making this research more focused and efficient.
Bjarne Knudsen
Chief Scientific Officer
CLC bio A/S
Aarhus, Denmark
|
Return to Emerging Research Fronts | Return
to Special Topics main menu
|