Ge Lab
Biopharmaceutical Discovery
Target Modulating Biologics
The development of monoclonal antibodies (mAbs) has been substantially advanced through the introduction of hybridoma technology in the 1970s and phage display methodologies in the 1990s, leading to the emergence of numerous complementary technologies for mAb discovery and optimization. These innovations have established mAbs as the predominant therapeutic modality within the biopharmaceutical sector. Nevertheless, the generation of therapeutic mAbs targeting challenging antigens remains a formidable undertaking. Notable examples of these particularly demanding objectives encompass (1) antibodies that demonstrate not merely binding capacity but also execute therapeutic functions, including protease inhibition and comprehensive viral neutralization; (2) mAb-based modulators of membrane-resident proteins such as G-protein coupled receptors and ion channels; and (3) biological therapeutics capable of traversing the blood-brain barrier (BBB).
Traditional protein engineering methods include directed evolution and computational designs based on structural biology and biophysical calculation. After decades of development, both methods have reached their limits showing obvious drawbacks – experiments can only survey a tiny fraction of entire sequence space, while computation often lack the accuracy needed for precise calculation. Recent advances in deep learning render us the third pillar on this task. To date, the data driven methods have been largely successful in structure prediction, but not fully utilized in biologics discovery. To revolutionize protein engineering to the next new level, I believe the synergies between experiment, computation, and deep learning can bring breakthroughs.
Funding
