Prospective identification of parasitic sequences in phage display screens. Issue 3 (9th November 2013)

• Record Type: Journal Article
• Authors: Matochko, Wadim L.; Cory Li, S.; Tang, Sindy K.Y.; Derda, Ratmir
• Is Part Of: Nucleic acids research. Volume 42:Issue 3(2014)
• Page Start: 1784
• Page End: 1798
• Publication Date: 2013-11-09
• Subjects: Nucleic acids -- Periodicals; Molecular biology -- Periodicals; 572.805
• DOI: 10.1093/nar/gkt1104
• Languages: English
• ISSNs: 0305-1048

Abstract

Abstract: Phage display empowered the development of proteins with new function and ligands for clinically relevant targets. In this report, we use next-generation sequencing to analyze phage-displayed libraries and uncover a strong bias induced by amplification preferences of phage in bacteria. This bias favors fast-growing sequences that collectively constitute <0.01% of the available diversity. Specifically, a library of 10 9 random 7-mer peptides (Ph.D.-7) includes a few thousand sequences that grow quickly (the 'parasites'), which are the sequences that are typically identified in phage display screens published to date. A similar collapse was observed in other libraries. Using Illumina and Ion Torrent sequencing and multiple biological replicates of amplification of Ph.D.-7 library, we identified a focused population of 770 'parasites'. In all, 197 sequences from this population have been identified in literature reports that used Ph.D.-7 library. Many of these enriched sequences have confirmed function (e.g. target binding capacity). The bias in the literature, thus, can be viewed as a selection with two different selection pressures: (i) target-binding selection, and (ii) amplification-induced selection. Enrichment of parasitic sequences could be minimized if amplification bias is removed. Here, we demonstrate that emulsion amplification in libraries of ∼10 6 diverse clones prevents the biased selection of parasitic clones.