Until now, we have assumed that just 20 amino acids form the building blocks of proteins; these building blocks are called proteinogenic amino acids. There are actually two additional proteinogenic amino acids, called selenocysteine and pyrrolysine, which are incorporated into proteins by special biosynthetic mechanisms (see"DETOUR: Selenocysteine and Pyrrolysine" at Cogniterra or in the print companion). Yet in addition to the 22 proteinogenic amino acids, NRPs contain non-proteinogenic amino acids, which expand the number of possible building blocks for antibiotic peptides from 20 to over 100.

Enlarging the amino acid alphabet spells trouble for our current approach to cyclopeptide sequencing. Indeed, the correct peptide now must “compete” with many more incorrect ones for a place on the leaderboard, increasing the chance that the correct peptide will be cut along the way.

For example, although Tyrocidine B1 contains only proteinogenic amino acids, its close relative, Tyrocidine B (Val-Orn-Leu-Phe-Pro-Trp-Phe-Asn-Gln-Tyr), contains a non-proteinogenic amino acid called ornithine (Orn). Because so many non-proteinogenic amino acids exist, bioinformaticians often assume that any integer between 57 and 200 may represent the mass of an amino acid; the “lightest” amino acid, glycine (Gly), has mass 57 Da, and most amino acids have masses smaller than 200 Da.

Exercise Break Applying LeaderboardCyclopeptideSequencing on the extended amino acid alphabet (i.e., every integer between 57 and 200 inclusively) to Spectrum₁₀ with N = 1000 returns 34 different linear peptides of maximum score. What are they? Solve this on Cogniterra.

Note: This exercise is difficult because it may take a long time to run if your solution is inefficient. Make sure you obtain an answer before clicking to start solving.

When we apply LeaderboardCyclopeptideSequencing for the extended alphabet to Spectrum₁₀, one of the highest-scoring peptides is VKLFPWFNQXZ, where X has mass 98 and Z has mass 65. Apparently, non-standard amino acids successfully competed with standard amino acids for the limited number of positions on the leaderboard, resulting in VKLFPWFNQXZ winning over the correct peptide VKLFPWFNQY. Since LeaderboardCyclopeptideSequencing fails to identify the correct peptide even with only 10% false and missing masses, our stated aim from the previous section is now even more important. We must determine the amino acid composition of a peptide from its spectrum so that we may run LeaderboardCyclopeptideSequencing on this smaller alphabet of amino acids.

STOP and Think How can we determine which amino acids are present in an unknown peptide using only an experimental spectrum?

From 20 to More than 100 Amino Acids