Whilst antibodies are produced by our immune cells to combat viruses and other pathogens, for a few decades, the use of biotechnology to produce antibodies as drugs is been in use for few decades now. This is because antibodies are exceptionally good at attaching, in particular to molecular structures based on lock-and-key principle. The use of antibodies is wide ranging from oncology to treatment of autoimmune disorders and neurodegenerative diseases.
However, to develop antibody drugs is a herculean task. The basic requirement to produce antibody drugs is to attach to its target molecule in an optimal manner. At the same time, the antibody drug must satisfy a number of additional criteria. For example, it should not set off an immune response in the body, should be efficient to produce drugs using biotechnology and should display stability over a long period of time.
In fact, once scientists discover that an antibody binds to the desired molecular structure that is the start of the development process, which is far from over. It is instead the start of the development process, wherein researchers use bioengineering to try to improve the properties of antibodies. To establish this, a team of researchers used machine learning method that supports optimization of development phase of antibodies, thereby helping to create more effective antibody drugs.
Meanwhile, when researchers optimize a complete antibody molecule in its therapeutic form, they start with an antibody lead candidate that joins reasonably well to the desired target structure. With this setup, researchers randomly mutate genes that carry the blueprint for the antibody in order to generate a few thousand associated antibody candidates in the lab.