Researchers used to start with an antibody lead candidate that binds relatively well to the required target structure while optimizing an entire antibody molecule in its therapeutic form.
FREMONT, CA: It is a well known fact that antibodies are produced by immune cells in the human body to protect it from viruses and other pathogens. However, with the technological advancement in biotechnology, many biotech firms are producing antibodies as drugs that are used by the pharmaceutical industry to treat patients with certain illnesses. This is because antibodies are able to bind to molecular structures as per the lock-and-key principle. Their use varies from oncology to the treatment of autoimmune diseases and neurodegenerative conditions.
However, developing such antibody drugs is not as simple as it may sound. The fundamental requirement is for an antibody to bind to its target molecule in an effective way. Simultaneously, an antibody-drug should meet a slew of additional criteria.
After scientists have discovered that an antibody can be bound to the desired molecular target structure, the development process is still under process. Rather, this is the start of a process in which researchers can try to enhance the antibody's properties using bioengineering.
Robots Are Unable to Manage More Than a Few Thousand
Researchers used to start with an antibody lead candidate that binds relatively well to the required target structure while optimizing an entire antibody molecule in its therapeutic form. The researchers then randomly mutate the gene that contains the antibody's blueprint in order to generate a few thousand antibody candidates in the lab. The next move is to look at them all to see which ones bind to the target structure the best.
Researchers used a CRISPR mutation method they produced a few years ago to produce around 40,000 similar antibodies based on the DNA sequence of the Herceptin antibody. Ten thousand of them bound well to the target protein in question, a particular cell surface protein, according to experiments. The scientists trained a machine learning algorithm using the DNA sequences of these 40,000 antibodies.
The qualified algorithm was then used to scan a database of 70 million possible antibody DNA sequences. The algorithm predicted how well the corresponding antibodies would bind to the target protein for each of the 70 million candidates, leading to a list of millions of sequences that should bind.
See Also: Top MedTech Companies