A comparative analysis of sequence composition in different lots of a phage display peptide library during amplification

Abstract

Background: To develop efficient selection strategies and improve the discovery of promising ligands, it is highly desirable to analyze the sequence composition of naïve phage display libraries and monitor the evolution of their peptide content during successive rounds of amplification. In the current study, we performed a comparative analysis of the compositional features in different lots of the same naïve phage display library and monitored alterations in their peptide compositions during three rounds of amplification.

Methods: We conducted three rounds of duplicate serial amplification of two different lots of the Ph.D.™-12 phage display library. DNA from the samples was subjected to Next-Generation Sequencing (NGS) using an Illumina platform. The NGS datasets underwent a variety of bioinformatic analyses using Python and MATLAB scripts.

Results: We observed substantial heterogeneity in the sequence composition of the two lots indicated by differences in the enhanced percentage of wildtype clones, reduced diversity (number of unique sequences), and increased enrichment factors (EFs) during amplification as well as by observing no common sequence between lots and decreased number of common sequences between the naïve library and the consecutive rounds of amplification for each lot. We also found potential propagation-related target-unrelated peptides (TUPs) with the highest EFs in the two lots, which were displayed by the fastest-propagating phage clones. Furthermore, motif analysis of the most enriched subpopulation of amplified libraries led to the identification of some motifs hypothesized to contribute to the increased amplification rates of the respective phage clones.

Conclusion: Our results highlight tremendous heterogeneity in the peptide composition of different lots of the same type of naïve phage display library, and the divergent evolution of their compositional features during amplification rounds at the amino acid, peptide, and motif levels. Our findings can be instrumental for phage display researchers by bringing fundamental insights into the vast extent of non-uniformity between phage display libraries and by providing a clear picture of how these discrepancies can lead to different evolutionary fates for the peptide composition of phage pools, which can have profound impacts on the outcome of phage display selections through biopanning.

Keywords: Amplification; Bias; Compositional heterogeneity; Enrichment factor; Illumina sequencing; Motif discovery; NGS; Peptide library; Phage display; Propagation-related peptides.

Fig. 1

Overview of the percentage of wildtypes (black) and the number of unique reads (blue) in the two lots in the naive library and through the different rounds of amplification. SA1: Serial amplification experiment 1 and SA2: Serial amplification experiment 2

Fig. 2

Overview of how the merged files representing sequences from different replicates were generated and how the generated files were used for different analyses. The figure was created with BioRender.com

Fig. 3

Stacked bar plots visualizing the frequency distribution of sequences observed in NGS data from the SA1 and SA2 experiments. Peptide sequences were grouped into bins (shown by different colors) according to their copy number. c.p.: copy number

Fig. 4

Illustration of the number of common sequences between the naïve library and different rounds of amplification for SA1 and SA2. In addition to the number of common sequences between the different datasets, the related percentages of sequences are also stated in parentheses

Fig. 5

Distribution of enrichment factors (EFs) during the different rounds of amplification (Round 1, 2, and 3) for different lots of the peptide phage display library (SA1 and SA2). The horizontal axis represents the EF of peptides and the vertical axis represents the number of peptides having the given EF

Fig. 6

Global frequency for the top 1000 peptide sequences ranked according to EF for both SA1 (upper graph) and SA2 (lower graph). The global frequencies are shown for the expected (theoretical) frequencies of amino acids as well as for the observed frequencies of amino acids in the naïve library and the three rounds of amplification (Round 1, Round 2, and Round 3)

Fig. 7

Positional frequency for the top 1000 sequences ranked according to EF for both SA1 (upper row) and SA2 (lower row). For each amino acid in the 12 different positions, the difference from the expected value was calculated and determined as a fraction. Red: overrepresented compared to the expected value, blue: underrepresented compared to the expected value

Fig. 8

The discovered hexapeptide motifs from the STREME analysis on the top 18,622 sequences (ranked based on EF) of the naïve library (NL), round 1 (R1), round 2 (R2), and round 3 (R3) of both SA experiments. The discovered tripeptide motifs can be seen in Supplementary Fig. A1