ARE FUNCTIONAL PROTEIN SEQUENCES RANDOM? Henrik Haraldsson Master Thesis Advisor: Carsten Peterson Abstract The question of whether proteins originate from random sequences of amino acids is addressed. A statistical block analysis of the sequences in the SWISS-PROT data base is performed, assigning a hydrophobicity to each residue according a multi-letter scale. Deviations in different directions from randomness are found when studying the interior separately from the endpoints of sequences, and when studying sequence interiors with average hydrophobicity separately from those with a large or small hydrophobicity. The distribution of total hydrophobicity among the sequences is also found to be wider than for randomness. These results confirm previous work done with a binary hydrophobicity scale. New relations between different deviations from randomness when using various hydrophobicity scales are also observed. Volume as a scale gives no deviation from randomness, but a particularly strong signal is obtained when using a certain combination of hydrophobicity and volume scales. Chou-Fasman propensities $P_\alpha$ and $P_\beta$ for $\alpha$- and $\beta$ secondary structure are also found to give different deviations from randomness. February 1999