Methods for identifying proteins by using partial sequences.

Methods for the identification of a protein segment by using the information from partial sequence analyses are described. If the protein sequence is known, a segment can usually be identified with confidence through comparison with the data file of all known sequences when the identity and position of only seven amino acid residues (not necessarily contiguous) are known. Partial sequences are obtained from extremely sensitive microsequencing procedures. Tissue is incubated with amino acids, one or more of which are distinctively radiolabeled. Proteins of interest are isolated and a sequenator experiment performed to locate the positions of radioactivity in an NH(2)-terminal segment of approximately 30 residues. We derive and investigate an equation for the probability of finding a unique match to any pattern of radioactivity. From this we suggest a new strategy. In one incubation, several amino acids are labeled with each kind of isotope. The most information is contained in patterns in which approximately equal numbers of positions are occupied by residues distinguished by different labels (including no label). The amino acid composition of the segment will typically not be known in advance. Labeling residues expected to occupy 36% of the positions suffices for a 98% chance of success in uniquely characterizing any human segment. Such a strategy will permit the identification of most proteins from a single tissue incubation. The mathematical discussion is general and applies to any segment from a sequence and to sequences obtained by any method. Improved identification procedures should expedite the accumulation of information on the expression and function of proteins.