In Silico Anibody Affinity Maturation:
Antibodies comprise a class of proteins indispensable in mediating humoral
immune responses. Their key role in combating diseases means that there is
much interest in designing artificial immunoglobulins, and altering the behaviour
of those innate to the organism. Therefore, just after vaccines, antibodies are
the second largest category of biopharmaceuticals [1].
However there are very few computational methods to date which aid in
designing antibodies. One notable example is the method described by Lippow
et al. [2], where existing antibodies were modified in silico to achieve increase in
affinity ranging from ten to hundred fold with more than 60% selected antibodies
capable of stronger binding than wild-type.
In this project I will explore possible methods of antibody design using computational
methods. Possible directions are numerous but can be easily divided
between ab initio design and developing improvements to existing antibodies.
Ab initio design is very challenging and it might not be possible to produce
an accurate method that given an antigen would produce a high-affinity binding
antibody. Work in this direction would rather include strategies to design efficient
artificial antibody libraries that would act as a starting point for antibody
design.
Enhancing affinity of existing antibodies includes the selection of favourable
mutations, improvement in prediction of loop conformation due to mutations,
analysis of docking of antibody to antigen and calculation of the
affinity of such binding. Selection of mutations favourable for affinity maturation
will be explored statistically by mapping mutations between all existing
antibody structures. Existing tools for loop structure prediction (e.g.
RosettaAntibody [3], FREAD [4]) can be employed to accurately predict antibody
conformation after mutations were applied. Docking tools designed specifically
for antibodies (SnugDock [5]) will play a crucial role when it comes to estimating
binding energy of the antigen-antibody complex.
This doctoral project's main aim will be to produce a computational method
improving the binding affinity of antibodies for their target antigens. Method
improving existing antibodies would be a sufficient achievement. Analysis of
favourable mutations increasing the desired properties of antibodies and an at-
tempt to design artificial antibody libraries would also make good contributions
to the field of antibody design.
References
- [1] Brekke OH, Sandlie I, Therapeutic antibodies for human diseases at the
dawn of the twenty- rst century, Nat Rev Drug Discov, 2:52-62. (2003)
- [2] Lippow SM, Wittrup KD, Tidor B, Computational design of antibody affinity improvement beyond in vivo maturation, Nature Biotechnology 25:
1171 - 1176. (2007)
- [3] Sircar A, Kim ET, Gray JJ, RosettaAntibody: antibody variable region
homology modeling server, Nucleic Acids Research, 2009, Vol. 37, No. suppl2
W474-W479
- [4] Chooi Y, Deane CM, FREAD revisited: Accurate loop structure prediction using a database search algorithm, Proteins: Structure, Function, and
Bioinformatics, Vol. 78, No. 6. (2010), pp. 1431-1440.
- [5] Sircar A, Gray JJ, 2010 SnugDock: Paratope Structural Optimization during Antibody-Antigen Docking Compensates for Errors in
Antibody Homology Models. PLoS Comput Biol 6(1): e1000644.
doi:10.1371/journal.pcbi.1000644
i-Patch Web Service: Prediction of Inter Protein Contact sites:
