Phage Display and Peptide Libraries: How New Peptides Are Discovered

In plain English

Finding a peptide (a short chain of amino acids) that sticks tightly to a chosen protein is a search problem. There are far too many possible sequences to make and test one at a time.

Phage display solves this with a trick. A bacteriophage (a virus that multiplies inside bacteria) is engineered so that a short peptide is built into one of its coat proteins, while the DNA coding for that peptide sits inside the same particle ^[1]. Make millions or billions of these particles, each displaying a different sequence, and the result is a library.

The library is poured over an immobilised target protein. Particles that stick are kept, the rest are washed away, and the kept particles are grown again in bacteria. After a few rounds of this, the surviving DNA is sequenced, and the sequence reveals which peptides bound ^[2].

Sticking to something is not the same as doing something useful, so hits then have to be made again as pure chemical peptides and retested.

The short version

• Each phage particle displays one peptide and carries its gene.
• Selection by binding is called biopanning, done in rounds ^[2].
• Libraries can hold 10⁸ or more different sequences ^[3].
• A binding hit is a starting point, not a finished molecule ^{[16, 17]}.

Where phage display came from

In 1985 George Smith showed that a foreign DNA fragment could be inserted into gene III of a filamentous phage so that the encoded amino acids were displayed on the surface of the virus particle, which could still enter a bacterial host. These “fusion phage” could be enriched more than 1,000-fold over ordinary phage by affinity for an antibody directed against the displayed sequence ^[1].

Five years later, three groups independently turned that idea into random peptide libraries. Scott and Smith built an “epitope library”, a mixture of phage clones each displaying one peptide, and showed that tens of millions of short peptides could be surveyed for tight binding to a binding protein ^[2]. Cwirla and colleagues built a library of 3 × 10⁸ recombinants displaying hexapeptides at the N-terminus of the phage pIII protein ^[3]. Devlin and colleagues displayed about 2 × 10⁷ different 15-residue sequences on phage M13 and isolated streptavidin-binding peptides whose core consensus was His-Pro-Gln; binding of those phage to streptavidin was blocked by biotin, the natural ligand ^[4]. In the same year, McCafferty, Griffiths, Winter and Chiswell displayed complete antibody variable domains on phage and showed that rare phage, about one in a million, could be isolated by affinity chromatography ^[5].

The 2018 Nobel Prize in Chemistry was awarded one half to Frances H. Arnold “for the directed evolution of enzymes”, and the other half jointly to George P. Smith and Sir Gregory P. Winter “for the phage display of peptides and antibodies” ^[19]. The Royal Swedish Academy of Sciences noted that Winter applied phage display to the directed evolution of antibodies, and that the first medicine developed with the method, adalimumab, was approved in 2002 ^[20].

See From Discovery to Approved Medicine for what happens after a binder is found, and A Short History of Peptide Science for the wider timeline.

How biopanning works

Selection on an immobilised target is called biopanning, and it is carried out in rounds. A typical round has five steps ^{[6, 16]}:

Three or four rounds are usual. Cwirla and colleagues sequenced 51 clones after three rounds of panning and found that all 51 carried an N-terminal tyrosine and 48 carried glycine in the second position: a clear consensus emerging from a random starting pool ^[3].

The word “round” matters here. Each pass through selection and amplification is a discrete generation, and the composition of the library changes at every pass, not only through selection but through growth.

Library formats: what is displayed, and where

Filamentous phage offer two main display sites, and they behave differently.

Peptides can also be constrained rather than linear. A linear peptide of 7 to 15 residues samples many shapes in solution, which costs binding affinity. Cyclic and bicyclic formats reduce that flexibility:

• Disulfide-cyclised libraries place cysteines at fixed positions so the displayed peptide forms a loop.
• Chemically bridged bicyclic libraries. Heinis, Rutherford, Freund and Winter designed phage repertoires with three reactive cysteines separated by random residues and conjugated them with tris-(bromomethyl)benzene, generating repertoires with two peptide loops anchored to a mesitylene core ^[8]. The displayed entity is then part peptide, part small molecule.

Constrained formats connect to the chemistry covered in Peptide Modifications and Disulfide Bonds in Peptides.

How large can a library be?

Library size is limited by how the DNA gets into the system. Phage and bacterial systems need transformation into living cells; cell-free systems do not.

Mattheakis and colleagues noted that their cell-free polysome system had the potential to screen libraries three to six orders of magnitude larger than the biological display systems available at the time ^[13]. A review of display technologies lists ribosome display, mRNA display, CIS display and covalent antibody display among the fully in vitro alternatives, with advantages in library size, speed and the use of unnatural amino acids ^[15].

Other library methods worth knowing

One-bead-one-compound (OBOC). Lam and colleagues synthesised peptides on solid-phase beads so that each bead carried one sequence, then screened the beads for binding and sequenced the peptide from the positive bead: the “one-bead, one-peptide” approach ^[9]. The library is chemical, not biological, so non-natural residues are easy to include. The same solid-phase chemistry is described in How Peptides Are Made.

DNA-encoded chemistry. Brenner and Lerner proposed encoding each member of a large chemical library with a unique nucleotide sequence built up in parallel with the chemistry. After binding to a target, the genetic tag is amplified and sequenced to decode which structure bound ^[10]. The approach has since developed into a standard industrial screening format ^[24].

mRNA display. Roberts and Szostak generated covalent fusions between an mRNA and the peptide it encodes by translating synthetic mRNAs carrying puromycin at their 3′ end. The stable link between the information and the peptide allows one mRNA to be enriched from a complex mixture on the basis of its encoded peptide ^[11]. Because translation happens in vitro, the genetic code can be reprogrammed, which is how mRNA display is used to make macrocyclic libraries containing modified backbones and unnatural side chains ^[14].

Ribosome display. Hanes and Plückthun kept a folded protein and its encoding mRNA attached to the ribosome, then enriched a single-chain antibody fragment 10⁸-fold over five rounds of transcription, translation, affinity selection and PCR ^[12].

From hit to validated binder: binding is not function

The most common failure in library work is believing a sequence that was never a real binder.

Target-unrelated peptides (TUPs). Reviews of biopanning describe two families of false positives. Selection-related TUPs bind something other than the target: the solid phase, the capturing reagent, the blocking agent or a contaminant in the target preparation. Propagation-related TUPs win because they grow faster in bacteria, and amplification between rounds makes a large contribution to their enrichment ^{[16, 17]}. Several sequences have been published repeatedly as target binders without being recognised as target-unrelated ^[16].

Amplification narrows the library. Deep sequencing of phage libraries showed that amplification of a model library of 10⁶ unique clones enriched about 150 clones, which came to dominate roughly 20% of the library. The authors concluded that screens based on repeated amplification plus small-scale sequencing identify a few binding clones and miss thousands of others ^[18].

Affinity is usually modest at first. In the 1990 hexapeptide selection, six chemically synthesised peptides from the selected set had binding affinities ranging from 0.35 µM to 8.3 µM, compared with 7.1 nM for a previously known high-affinity ligand for the same antibody ^[3]. A hit from a naive library is typically a micromolar starting point that needs further rounds of mutagenesis and chemistry.

A short checklist for reading a library paper:

1. Is the hit resynthesised? A phage clone that binds is weaker evidence than a pure synthetic peptide that binds.
2. Is there a concentration–response curve with a reported Kd, IC50 or EC50, and under what assay conditions? See EC50, IC50, Kd and Ki.
3. Are negative controls shown for the plate, beads, blocking agent and capture reagent ^[16]?
4. Is the peptide checked against known TUP lists and against the possibility of propagation bias ^[17]?
5. Is binding linked to any measured function in a defined system, or only to binding? Occupying a site does not say what happens next; see How Peptide Signalling Works.

What this means in Australia

• Genetically modified organisms are regulated. Phage display uses engineered bacteriophage and bacterial hosts. In Australia, work with genetically modified organisms is a “GMO dealing” under the gene technology regulatory scheme, and dealings are classified: some are exempt or notifiable low-risk, while others require a licence from the Gene Technology Regulator, including dealings not involving intentional release (DNIR) that do not meet the criteria for the lower classifications ^[23].
• A discovery sequence is not an approved product. A peptide identified in a library screen has no regulatory standing as a medicine. It is not entered on the Australian Register of Therapeutic Goods (ARTG) and has not been assessed by the TGA ^[21].
• “Research use only” does not change status. The TGA has stated that a “research use only” disclaimer does not make supply lawful, change a product’s regulatory status, permit importation, or remove advertising or supply obligations ^[21].
• Scheduling depends on the substance. Several peptides are named in Schedule 4 (prescription only) and Appendix D, clause 5 (possession without authority is illegal) of the current Poisons Standard (October 2026) ^[22]. See Are Peptides Legal in Australia?.

This is general information, not legal advice. Laws change; check the current Poisons Standard and TGA guidance.

Frequently asked questions

What is phage display?

It is a method in which a peptide or protein is displayed on the surface of a bacteriophage while the DNA encoding it stays inside the same particle, so that selection by binding can be read back as a sequence ^{[1, 2]}.

What is biopanning?

Biopanning is affinity selection of a display library on an immobilised target, carried out in rounds of binding, washing, elution and amplification ^{[6, 16]}.

Who won the 2018 Nobel Prize in Chemistry for phage display?

The 2018 prize was divided, one half to Frances H. Arnold for the directed evolution of enzymes, and the other half jointly to George P. Smith and Sir Gregory P. Winter for the phage display of peptides and antibodies ^[19].

How many different peptides can a library contain?

Phage libraries of 10⁷–10⁸ distinct sequences are routine ^{[3, 4]}, and cell-free formats have been built at 10¹² ^[13]. The practical limit for phage and bacterial systems is set by getting DNA into living cells.

Why does a selected peptide often fail when it is resynthesised?

Because selection can reward things other than target binding, including binding to the plate or capture reagent and faster growth during amplification ^{[16, 17]}, and because affinities from naive libraries are often micromolar rather than nanomolar ^[3].

What is the difference between phage display and mRNA display?

Phage display is carried out in living bacteria and the tag is the phage genome. mRNA display is carried out in vitro, and the tag is an mRNA covalently fused to the peptide it encodes, which also allows unnatural residues ^{[11, 14]}.

References

1. Smith GP. Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science. 1985;228(4705):1315–1317. doi:10.1126/science.4001944. PMID: 4001944
2. Scott JK, Smith GP. Searching for peptide ligands with an epitope library. Science. 1990;249(4967):386–390. doi:10.1126/science.1696028. PMID: 1696028
3. Cwirla SE, Peters EA, Barrett RW, Dower WJ. Peptides on phage: a vast library of peptides for identifying ligands. Proc Natl Acad Sci U S A. 1990;87(16):6378–6382. doi:10.1073/pnas.87.16.6378. PMID: 2201029
4. Devlin JJ, Panganiban LC, Devlin PE. Random peptide libraries: a source of specific protein binding molecules. Science. 1990;249(4967):404–406. doi:10.1126/science.2143033. PMID: 2143033
5. McCafferty J, Griffiths AD, Winter G, Chiswell DJ. Phage antibodies: filamentous phage displaying antibody variable domains. Nature. 1990;348(6301):552–554. doi:10.1038/348552a0. PMID: 2247164
6. Ledsgaard L, Kilstrup M, Karatt-Vellatt A, McCafferty J, Laustsen AH. Basics of antibody phage display technology. Toxins (Basel). 2018;10(6):236. doi:10.3390/toxins10060236. PMID: 29890762
7. Kuzmicheva GA, Jayanna PK, Sorokulova IB, Petrenko VA. Diversity and censoring of landscape phage libraries. Protein Eng Des Sel. 2009;22(1):9–18. doi:10.1093/protein/gzn060. PMID: 18988692
8. Heinis C, Rutherford T, Freund S, Winter G. Phage-encoded combinatorial chemical libraries based on bicyclic peptides. Nat Chem Biol. 2009;5(7):502–507. doi:10.1038/nchembio.184. PMID: 19483697
9. Lam KS, Salmon SE, Hersh EM, Hruby VJ, Kazmierski WM, Knapp RJ. A new type of synthetic peptide library for identifying ligand-binding activity. Nature. 1991;354(6348):82–84. doi:10.1038/354082a0. PMID: 1944576
10. Brenner S, Lerner RA. Encoded combinatorial chemistry. Proc Natl Acad Sci U S A. 1992;89(12):5381–5383. doi:10.1073/pnas.89.12.5381. PMID: 1608946
11. Roberts RW, Szostak JW. RNA-peptide fusions for the in vitro selection of peptides and proteins. Proc Natl Acad Sci U S A. 1997;94(23):12297–12302. doi:10.1073/pnas.94.23.12297. PMID: 9356443
12. Hanes J, Plückthun A. In vitro selection and evolution of functional proteins by using ribosome display. Proc Natl Acad Sci U S A. 1997;94(10):4937–4942. doi:10.1073/pnas.94.10.4937. PMID: 9144168
13. Mattheakis LC, Bhatt RR, Dower WJ. An in vitro polysome display system for identifying ligands from very large peptide libraries. Proc Natl Acad Sci U S A. 1994;91(19):9022–9026. doi:10.1073/pnas.91.19.9022. PMID: 7522328
14. Josephson K, Ricardo A, Szostak JW. mRNA display: from basic principles to macrocycle drug discovery. Drug Discov Today. 2014;19(4):388–399. doi:10.1016/j.drudis.2013.10.011. PMID: 24157402
15. Galán A, Comor L, Horvatić A, et al. Library-based display technologies: where do we stand? Mol Biosyst. 2016;12(8):2342–2358. doi:10.1039/c6mb00219f. PMID: 27306919
16. Vodnik M, Zager U, Strukelj B, Lunder M. Phage display: selecting straws instead of a needle from a haystack. Molecules. 2011;16(1):790–817. doi:10.3390/molecules16010790. PMID: 21248664
17. Bakhshinejad B, Zade HM, Shekarabi HS, Neman S. Phage display biopanning and isolation of target-unrelated peptides: in search of nonspecific binders hidden in a combinatorial library. Amino Acids. 2016;48(12):2699–2716. doi:10.1007/s00726-016-2329-6. PMID: 27650972
18. Matochko WL, Chu K, Jin B, Lee SW, Whitesides GM, Derda R. Deep sequencing analysis of phage libraries using Illumina platform. Methods. 2012;58(1):47–55. doi:10.1016/j.ymeth.2012.07.006. PMID: 22819855
19. The Nobel Prize in Chemistry 2018: summary. NobelPrize.org. nobelprize.org. Accessed 25 October 2026.
20. Royal Swedish Academy of Sciences. The Nobel Prize in Chemistry 2018: press release. NobelPrize.org. nobelprize.org. Accessed 25 October 2026.
21. Therapeutic Goods Administration. Understanding your responsibilities when importing, compounding and supplying unapproved peptide products. Safety advisory, 13 April 2026. tga.gov.au. Accessed 25 October 2026.
22. Therapeutic Goods (Poisons Standard—October 2026) Instrument 2026 (F2026L01327). Federal Register of Legislation. legislation.gov.au. Accessed 25 October 2026.
23. Office of the Gene Technology Regulator. Types of GMO dealings. ogtr.gov.au. Accessed 25 October 2026.
24. Gironda-Martínez A, Donckele EJ, Samain F, Neri D. DNA-encoded chemical libraries: a comprehensive review with successful stories and future challenges. ACS Pharmacol Transl Sci. 2021;4(4):1265–1279. doi:10.1021/acsptsci.1c00118. PMID: 34423264