Slightly belated, these are our thoughts on the MABRA workshop at the University of Surrey, which five OPIGlegts attended in January 2020.
General thoughts
The MABRA workshop differed quite strongly from the larger conferences we had attended before. With about 50 attendees, the conference allowed for a very direct engagement with the computational antibody research field, a welcome change from the only other conference I (Constantin) had attended so far, which was the ISMB in Basel with several thousand attendees. As such, the conference allowed for talks to a very targeted audience as well as plenty of lively discussion in between talks.
We learned some really interesting things both in and out of talks. One of my (Eve) favorite new facts came from comparative immunology – chickens have a single V gene (humans have 36-49 functional V genes).
Thoughts on some specific talks
With the OPIG antibody research focus lying on the analysis of antibody structure, we want to highlight three of the talks given at the workshop that delved into the three dimensional aspect of antibody research.
Dr Victor Greiff, University of Oslo, Norway: Unraveling the rules of paratope-epitope interaction using network theory and machine learning
Dr Greiff presented several recent works in his group, particularly interesting to me (Constantin) was his presentation of a previously published study on the ability to predict and categorize antibody-antigen interactions via a vocabulary of interaction motifs, which is analysed in the linked paper from a structural, machine learning and network perspective.
Professor Roland Dunbrack, Fox Chase Center, Philadelphia, USA: Structural bioinformatics of antibodies
Professor Dunbrack presented recent work in his group on clustering of antibody complementarity determining regions, revisiting the 2011 canonical classes [1] with new data and some additional structural analysis (e.g. correcting for peptide flips by looking at electron density). He further presented research recently made available as a preprint, in which he stipulates that the DE loop adjacent to CDR1 and CDR2 should be regarded as a fourth CDR and presents canonical families for this CDR type.
Professor Andrew Martin, University College London: Analysis and prediction of antibody sequence and structure for improved therapeutics
Professor Martin provided an overview of the antibody informatics tools and databases developed and curated by his group (available at http://www.bioinf.org.uk/abs/). Of particular note is the new antibody modelling capability on the website, abYmod. The website has also been updated recently making it super user friendly.
Conclusion
In summary, the MABRA workshop provided a great overview of many of the different subfields of adaptive immune receptor research. It was also a really kind and receptive audience for both of our first ever conference talks. Many thanks to the organisers and we hope to have the opportunity to attend the next one!
Constantin & Eve
[1] North, Benjamin, Andreas Lehmann, and Roland L. Dunbrack Jr. “A new clustering of antibody CDR loop conformations.” Journal of molecular biology 406.2 (2011): 228-256.
