A Voyage through Protein Sequence Space

The search for novel proteins is complicated by the diversity of amino acids, the non-linear relationship between amino acid sequence and folding and the huge size of sequence space. The goal of this research is to discover patterns in amino acid composition and sequence that relate to known or novel folds, thereby helping to constrain the search space for new functional proteins. This approach is referred to as sequenomics. First, in the construction phase, we demonstrate how sequence space can be meaningfully described and what kind of algorithms and sequence characteristics can be taken into account. Second, in the analytic phase, we describe characteristics of natural and artificial sequences in the sequence space. We find that a coordinate system based on substitution rates can reveal hidden patterns for α-helices and β-sheets in biological sequences. Furthermore, multiple local variance measurements suggest that biological sequences have a complex information signature that distinguishes them from randomly generated sequences. These observations pose new hypotheses about the protein sequence-structurefunction relationship that can help to focus the search for new functional proteins.