In total there are 20 different amino acids that have different functions such as digesting food, assisting in the growth and repair of tissues, or forming the immune system that defends us from external pathogens.
Each protein is a sequence of these amino acids placed one behind the other, depending on the number of them and their order, the different proteins with different functions are obtained.
Each necessary amino acid is encoded in the RNA using groups of three bases called codons and formed by the famous letters CAGT with the exception of T (Thymine), which in the RNA is replaced by a U (Uracil). (eg CGU >>> Arginine)
With these instructions, the RNA leaves the nucleus and once in the cytoplasm of the cell, the ribosomes are in charge of manufacturing the proteins using the amino acids that "float" in the cytoplasm, in what is known as "Protein synthesis".
Scientists have been trying for a long time to design and manufacture artificial proteins that could replace natural ones, but as you may have deduced by now from the post, that is extremely difficult for a mortal.
The use of machine learning combined with the study of huge amounts of data applying big-data can reveal the rules of the basic structures of protein formation and design artificial proteins.
This would undoubtedly have many and wonderful applications in the field of health, being able to manufacture proteins that replace those that do not work correctly in the cells of our body.
More information:
https://phys.org/news/2020-07-machine-reveals-recipe-artificial-proteins.html