NaturalAntibody was founded by Konrad Krawczyk who spent the last ten years studying computational approaches to improve the design of antibody-based therapeutics. Work in close collaboration within leading and academic institutions and pharmaceutical companies led to uncovering multiple avenues in which antibody data can be employed to create advanced computational models holding promise of accelerating therapeutic antibody design.

Such efforts laid the foundation for NaturalAntibody that focuses its resources on collecting, generating and analysing antibody data with the ultimate aim of end-to-end computational discovery of novel biotherapeutics.

Our expertise and interest in antibodies

Antibodies are proteins of the immune systems tasked with recognizing noxious molecules for elimination. As a result of this function, antibodies have been harnessed as biotherapeutics with an ever expanding presence in the clinic. Successful exploitation of these molecules for therapeutic purposes requires ever-deepening our understanding of the biological features of these molecules that make them into versatile binders. Using computational methods to achieve this task hold a promise in unlocking the new depths of the therapeutic potential of antibodies, saving on time and resources in delivering novel drugs to clincal use.Computational methods have matured enough for facilitating therapeutic antibody development and are currently employed by leading pharmaceutical companies. More than thirty years of therapeutic antibody development produced plentiful data on biological and therapeutic activity of antibodies. Such rich data resources can now be used to deliver on the ultimate promise of computational antibody discovery by creating advanced statistical models to design, analyze and optimize novel antibody therapeutics. Via repurposing and generating new data NaturalAntibody develops computational tools that aim to greatly reduce the time and cost of development of novel antibody therapeutics.

Computational methods are currently mature enough to provide value to antibody discovery pipelines, reducing the cost and time in therapy delivery to the clinic. The fundamental element in developing such methods is data.

After thirty years of monoclonal antibody therapy development there is plentiful data availalbe in the public domain alone. Such data is however scattered across diverse sources, such as patents, scientific publications, sequence/structure repositories and others.

We are re-purposing publicly available information and generating our own data to draw a wholesome picture of antibody biology and engineering capacity. Ingesting such data into our computational models develops the necessary bioinformatics capacity to tackle important issues in antibody engineering, accelerating monoclonal antibody therapy delivery to the clinic.


We publish our research results with many collaborators across academia and industry.
Our collaborators include:


Oxford UniversityCambridge UniversityUtrecht UniversityDanish Technical University


AstraZenecaBoehringer IngelheimUCB Pharma