Computational methods for predicting phenotypic effects of mutations using large sequence and deep mutational scan data

Abstract

Proteins are very important biological macromolecules that are involved in many cellular processes. Transport molecules such as haemoglobin, antibodies which are related to immune response, enzymes which catalyze chemical reactions, and structural matrices such as keratin or collagen are all proteins. In an apparently striking recursive role, all proteins are synthesized by other special proteins called ribosomes.[1] Thus any quest for understanding the basic cellular or disease biology mostly narrows the search to activity of some proteins or their failure. Whether it is non-communicable diseases such as cancers or Alzheimer’s or communicable diseases with bacterial or viral infections, the fundamental interest is always in knowing what went wrong with the most important proteins in healthy cells or how to block the bacterial or viral proteins from performing their expected functions.[2–4] Chemically speaking, proteins are polypeptide chains, linear polymers of amino acids connected by peptide bonds.[5] They are synthesized by ribosomes, with combinations of the 20 naturally occurring amino acids, appearing in a unique sequence. The linear polypeptide chain, which is also known as the primary structure of the protein, undergoes further physical transformations and levels of organization before it becomes functional.[6] Driven by non-bonded interactions such as hydrogen bonds or salt bridges among the amino acids which are either near or distal in sequence, proteins achieve secondary structures such as helices, β-sheets, turns or coils, which are further organized into a tertiary folded structure.[6] The ability to form different secondary structures varies for the side chains of different amino acids.[7] An even further organization into a quaternary structure happens with proteins which are multimeric in nature.Proteins are very important biological macromolecules that are involved in many cellular processes. Transport molecules such as haemoglobin, antibodies which are related to immune response, enzymes which catalyze chemical reactions, and structural matrices such as keratin or collagen are all proteins. In an apparently striking recursive role, all proteins are synthesized by other special proteins called ribosomes.[1] Thus any quest for understanding the basic cellular or disease biology mostly narrows the search to activity of some proteins or their failure. Whether it is non-communicable diseases such as cancers or Alzheimer’s or communicable diseases with bacterial or viral infections, the fundamental interest is always in knowing what went wrong with the most important proteins in healthy cells or how to block the bacterial or viral proteins from performing their expected functions.[2–4] Chemically speaking, proteins are polypeptide chains, linear polymers of amino acids connected by peptide bonds.[5] They are synthesized by ribosomes, with combinations of the 20 naturally occurring amino acids, appearing in a unique sequence. The linear polypeptide chain, which is also known as the primary structure of the protein, undergoes further physical transformations and levels of organization before it becomes functional.[6] Driven by non-bonded interactions such as hydrogen bonds or salt bridges among the amino acids which are either near or distal in sequence, proteins achieve secondary structures such as helices, β-sheets, turns or coils, which are further organized into a tertiary folded structure.[6] The ability to form different secondary structures varies for the side chains of different amino acids.[7] An even further organization into a quaternary structure happens with proteins which are multimeric in nature.[8]