Peptide Library Design and Synthesis

A peptide library is a set of related sequences made so that a screen can assign activity to position, length, or substitution. The design of that set is the scientific product. Synthesis is the means. A 96-well plate of random peptides is not a library in this sense; it is a pile. A well-designed alanine scan of a 14-mer is a library even if it contains only fourteen members.

ChinaPeptides manufactures research libraries under peptide libraries, using the same chemical constraints as custom peptide synthesis. This article walks through common designs, what mixture synthesis can and cannot claim, and how to avoid screens that cannot be deconvoluted.

Write the question the plate must answer

Which residues of this epitope contact the antibody? What is the minimal active fragment? Which substitution improves protease resistance in vitro without killing binding? Each question implies a different matrix. If you cannot state the question, you will over-build the library and under-power the assay.

Throughput of the assay sets the maximum useful size. A low-throughput SPR run cannot consume a 10,000-member mixture with a straight face. A plate-based ELISA can. Design to the readout, then to chemistry.

Overlapping fragment libraries

A protein region is cut into peptides of fixed length (often 12–20 residues) with a defined overlap (often 4–10 residues). Hits map to a window; the overlap is what makes the map readable. Too little overlap and a split epitope is missed. Too much overlap and you pay for redundant wells.

Termini should be consistent across the set (commonly acetyl/amide for internal stretches). A single biotinylated or Cys-terminated version of each peptide is a different library for immobilization; do not mix handle-bearing and handle-free peptides in one “hit list” without noting the handle.

Alanine and positional scanning

Alanine scanning replaces each non-alanine residue in turn. It is discrete, interpretable, and small. It misses substitutions that need a larger side chain. Positional scanning libraries (one position varied across many amino acids, others fixed or as mixtures) explore more chemistry per well but require a deconvolution plan. If a position is a mixture of 19 amino acids, a hit says “this position matters” more loudly than it says “use arginine.”

D-amino acid or N-methyl scans are specialized SAR libraries. They belong with unnatural amino acid planning because coupling efficiency will not be uniform across the set. Uneven coupling in a mixture well biases the actual composition away from the intended equimolar design.

Library type	What a hit means	Main design trap
Overlapping fragments	Activity lives in this window	Overlap too small; inconsistent termini
Alanine scan	This side chain contributes	Ala is not a neutral deletion of volume
Positional scanning (discrete)	This residue at this position works	Plate too large for the assay
Positional scanning (mixture)	This position is sensitive	Unequal coupling; no deconvolution path
Truncation series	Minimal length	Losing a charge that was only for solubility

Mixture synthesis and honesty about composition

Split-and-pool or premixed amino-acid couplings can generate many sequences in one vessel. The crude is a population. HPLC purity of “the library” is not a single-peptide percent. QC is statistical: mass envelopes, a subset of wells sequenced or individually resynthesized, and coupling protocols meant to equalize reactivity. You cannot treat a mixture well as a 95% pure compound.

For hit confirmation, resynthesize the implied discrete peptide at a real purity grade. The library finds candidates. The single peptide is the reagent you publish. Budget that confirmation synthesis from the start.

Overlapping fragment sets have a quality problem that mixture libraries do not: silent truncations at different wells look like a clean length map. If well 12 is a deletion of the C-terminal four residues, you may call a false minimal epitope. A sampling LC-MS of every nth well, plus mandatory MS on every hit well before you redesign the protein, is cheap compared with a wrong conclusion. Encode the intended full length in the well file and refuse to interpret a well whose observed mass is not the designed peptide.

Purity, scale, and plate logistics

Screening libraries are often made at lower purity and small scale (sub-milligram to a few milligrams). That is acceptable when the screen is comparative inside the set and every well is made the same way. It is unacceptable when one “hit” is a 50% peptide and the neighbors are 90%, and you rank them as SAR. If ranking matters, pay for discrete, characterized wells or normalize by peptide content.

Difficult sequences inside a library will fail silently—empty wells or truncated material—unless QC catches them. Flag known hard motifs so those wells can be made with the tactics in difficult peptide synthesis or omitted with a note. A missing well is better than a well of the wrong molecule labeled as the designed sequence.

Handles and library-wide modifications

If every peptide needs an N-terminal biotin for a streptavidin plate, that handle is part of the design, not an afterthought on half the plate. The same applies to a C-terminal Cys for conjugation. Document it in every sequence file you send. Modification chemistry is described generally under peptide modification; at library scale, simplicity wins. Fancy dyes on 400 wells are a different project from a few fluorescent peptides.

How to send a library request

Provide a table: well identifier, sequence, termini, handles, and any well-specific notes. State the assay, the preferred scale, and whether mixtures are allowed. State the confirmation policy (resynthesize hits at ≥95%). Use request a quote with the table attached. If you only have a protein sequence and an overlap rule, say so and ask for a proposed fragment list before synthesis starts.

Libraries are research tools. A screen hit is not a development candidate by itself.