Whole protein sequencing and quantification without proteolysis, terminal residue cleavage, or purification: A computational model

Abstract

Sequencing and quantification of whole proteins in a sample without separation, terminal residue cleavage, or proteolysis are modeled computationally. Similar to recent work on DNA sequencing (PNAS 113, 5233-5238, 2016), a high-volume conjugate is attached to every instance of amino acid (AA) type AAi, 1 [≤] i [≤] 20, in an unfolded whole protein, which is then translocated through a nanopore. From the volume excluded by 2L residues in a pore of length L nm (a proxy for the blockade current), a partial sequence containing AAi is obtained. Translocation is assumed to be unidirectional, with residues exiting the pore at a roughly constant rate of [~]1/s (Nature Biotechnology 41, 1130-1139, 2023). The blockade signal is sampled at intervals of 1 s and digitized with a step precision of 70 nm3; the positions of the AAis are obtained from the positions of well-defined quantum jumps in the signal. This procedure is applied to all 20 standard AA types, the resulting 20 partial sequences are merged to obtain the whole protein sequence. The complexity of subsequence computation is O(N) for a protein with N residues. The method is illustrated with a sample protein from the human proteome (Uniprot id UP000005640_9606). A mixture of M protein molecules (including multiple copies) can be sequenced by constructing an M x 20 array of partial sequences from which proteins occurring multiple times are first isolated and their sequences obtained separately. The remaining M singly-occurring molecules are detected from M disjoint paths through the 20 columns of the reduced M x 20 array. Detection complexity is O(M20), which is nominally in polynomial time but practical only for small M; to use this method a sample may be subdivided into subsamples down to this level. Quantification of proteins can be done by sorting their computed sequences on the sequence strings and counting the number of duplicates. The possibility of translating this procedure into practice and related implementation issues are discussed.