Table 1 Summary of libraries and triage results for lead identification.

From: A multi-faceted discovery strategy identifies functional antibodies binding to cysteine-rich domain 1 of hDKK1 for cancer immunotherapy via Wnt non-canonical pathway

Project

Library

No. clones reformatted for primary assays

TCF/LEF functional assay hits

Primary immune cell activators

Tumor killer

SC-52

VHH hShuffle/Hyperimmune

113

5

4

4

TB643

ML Approach

97

12

8

6

(VHH hShuffle/HI)

KNN based classifier

TB758

ML Approach

170

7

7

5

(VHH hShuffle/HI)

CNN based neural network

  1. The VHH hShuffle and hShuffle HI library series have a partially humanized framework which lowers the immunogenicity of the sequences isolated from the library for drug development. Specifically, framework 1, 3, and 4 are humanized. Framework 2 is a camelid framework to allow for better stability of the library. With the help of our in-house oligo printing at Twist, we have each CDR sequence individually synthesized instead of using degenerate codons. Both of these libraries are combinatorically generated which gives us a diversity of ~e9–e10. The hShuffle library has natural llama CDR sequences for CDR 1, 2, and 3. The hShuffle HI library has llama CDR sequences for CDR 1 and 2 and human CDR sequences for CDR3. This increases the diversity of CDR3 ~ 2.5 million sequences. These 2.5 million sequences are derived from human naive and memory B-cell sequences. This significantly increases the diversity of the library. The unique sequences from phage ELISA screening were selected for reformatting, production, and assay characterization.
Back to article page