A novel sequential heavy-chain (HC) preselection and remating workflow implemented in yeast surface display technology allows researchers to reallocate finite library capacity, prioritizing light-chain (LC) sampling depth over broad heavy-chain diversity. Combinatorial fragment antigen-binding (Fab) libraries typically encode heavy–light chain pairing spaces that can exceed 10^14 potential combinations, a scale that far surpasses the 10^7 to 10^9 diversity limits of eukaryotic display systems. Because direct Fab screening methodologies must commit both the heavy and light chain dimensions simultaneously, designers have traditionally faced an inherent constraint in how finite library capacity is distributed between these two domains. This new staging approach addresses that structural limitation by decoupling the selection process, effectively shifting the distribution of sampling opportunities.
The research team demonstrated this method through a sequential workflow in which antigen-compatible HC variants are first enriched as haploids and subsequently remated with a complete light-chain repertoire. By using a count of 18,023 sorted cells as an upper bound for the staged HC variants, the researchers calculated the nominal LC pairing opportunities available per HC. In conventional direct screening, these opportunities were approximately 16 per HC, whereas the staging workflow provided at least 7.8 × 10^3 pairing opportunities per retained HC. This represents an estimated 500-fold increase in the sampling depth for the light chains, achieved without requiring an expansion in the total number of unique HC–LC pairings that must be constructed and processed during the final screening phase.
A critical component of this workflow is that the HC-first filtering step operates exclusively on HC variants that remain both displayable and antigen-competent even in the absence of a partner light chain. During validation with a five-antibody reference panel, the team found that four display-competent HCs showed no detectable cognate-antigen binding when tested in the HC-only format at a concentration of 100 nM, or 500 nM for the CD40 target.
In a practical application targeting the SARS-CoV-2 spike protein, the researchers compared the staging workflow against direct Fab screening using the same starting repertoires. They observed that the variety of HC clonotypes among sampled clones decreased from 15 to 4 throughout the staged process. Within the staged samples, the dominant clonotype, CLR201, accounted for 60.9% of the clones, exhibiting 12 distinct LC complementarity-determining region patterns among its 14 sampled clones. By contrast, direct Fab screening from the same repertoires returned 8 clonotypes with a dominant frequency of 30% (23 and 10 clones sampled, respectively).
The CLR201 clone, recovered through the staging process, bound to the receptor-binding domain (RBD) with an apparent on-yeast half-maximal effective concentration (EC50) of 2.75 nM. This value represents an improvement over the 3.70 to 9.73 nM range observed for the clones retrieved via direct Fab screening. Furthermore, when converted into a full-length IgG1 format, the CLR201 antibody demonstrated sub-nanomolar apparent EC50 values against both the wild-type RBD and five different RBD variants, and successfully neutralized the D614G spike-pseudotyped virus.
The study concludes that the reallocation of diploid library capacity toward LC sampling depth is a result of the library architecture itself, rather than an antigen-specific phenomenon. Whether this deeper LC sampling reliably yields superior binders remains dependent on the specific antigen and the constraints of the screening campaign. As a proof of principle, the two campaigns compared here were not stringency-matched; the direct screening approach maintained a broader representation of the initial heavy-chain pool, while the staging method prioritized depth, successfully isolating the clone with the lowest apparent on-yeast EC50 among those tested.
Subject of Research: Sequential heavy-chain preselection for optimizing Fab antibody library screening.
Article Title: Staged heavy-chain preselection reallocates Fab library capacity from heavy-chain breadth to light-chain sampling depth: a single-target, same-repertoire comparison with direct Fab screening
Article References: Kim, Y., Kwon, H., Hong, J., Kang, C. K., Park, W. B., Kim, H.-R., & Lee, C.-H. (2026). Staged heavy-chain preselection reallocates Fab library capacity from heavy-chain breadth to light-chain sampling depth: a single-target, same-repertoire comparison with direct Fab screening. Journal of Biological Engineering. https://doi.org/10.1186/s13036-026-00794-y
Image Credits: AI Generated
DOI: 10.1186/s13036-026-00794-y
Keywords: Antibody Discovery, Yeast Surface Display, Fab Libraries, Protein Engineering, Heavy-Chain Preselection, Light-Chain Sampling, Affinity Maturation, Combinatorial Libraries, SARS-CoV-2, Spike Protein, CLR201, Neutralizing Antibodies, Library Architecture, Molecular Biology, Biotechnology, Receptor-Binding Domain, Protein Pairing
Cite Scienmag News
Gregory Coleman. (November 17, 2026). Staged Preselection Enhances Light-Chain Sampling Depth in Fab Antibody Libraries. Scienmag. https://scienmag.com/staged-preselection-enhances-light-chain-sampling-depth-in-fab-antibody-libraries/
Gregory Coleman. "Staged Preselection Enhances Light-Chain Sampling Depth in Fab Antibody Libraries." Scienmag, 17 November 2026, https://scienmag.com/staged-preselection-enhances-light-chain-sampling-depth-in-fab-antibody-libraries/. Accessed 20 November 2026.
Gregory Coleman. "Staged Preselection Enhances Light-Chain Sampling Depth in Fab Antibody Libraries." Scienmag. November 17, 2026. https://scienmag.com/staged-preselection-enhances-light-chain-sampling-depth-in-fab-antibody-libraries/








