We are opening a new landscape of novel biologics by applying machine learning specifically to immunology, guided by our team of experts in biologics discovery and development.
To Parallelize is to optimize multiple relevant desired parameters simultaneously to rapidly modulate key drug-like attributes of protein therapeutics.
Traditional protein identification and optimization involves inefficient, independent sequential iteration and protracted combination steps.
Seismic’s IMPACT platform enables parallelized multi-property optimization by integrating machine learning with structural biology, protein engineering, and translational immunology for the accelerated generation of novel biologics.
The IMPACT platform opens new protein sequence space, allows the simultaneous optimization of multiple drug-like properties, including immunogenicity, and reduces both the number and duration of design/test cycles.
New Sequence Space
The IMPACT platform allows us to find previously unexplored sequences and to design proteins with optimal drug-like properties.
Invisibilization
The IMPACT platform allows us, using our Invisi-B and Invisi-T ML tools, to remove B and T cell epitopes from novel biologic drugs to make them invisible to the immune system while retaining fitness and function of these drugs.
Design Test Cycle
The IMPACT platform enables parallelization, thereby removing many cycles of iteration and mutation combinations and shortening the time of each cycle to rapidly identify top candidates.
Our pipeline programs are derived from our IMPACT platform. Our proven team of integrated protein engineers and immunologists work collectively with our cutting-edge machine learning team to discover and develop novel enzyme and antibody drug candidates for autoimmune diseases.
The following explains how we’ve applied our IMPACT platform and integrated machine learning into our discovery efforts to make improved drug candidates.
We leveraged the IMPACT platform to screen thousands of naturally occurring enzymes that modify pathogenic antibodies, and, driven by machine learning, rapidly engineered them into novel enzymes with ideal drug-like characteristics using multi-parameter optimization. Using this approach, we have discovered and developed invisibilized, developable, and functional novel pan-IgG proteases.
We leveraged the IMPACT platform to discover and develop DcB antibodies that simultaneously engage multiple inhibitory pathways in more than one immune cell type to regulate both sides of the immune cell synapse. In particular, machine learning was critical to designing novel inhibitory Fc gamma receptor IIb (FcγRIIb)-selective Fc domains by predicting mutations that tune for the desired binding properties at the Fc/FcγRIIb interface. Our ML approach was also instrumental in predicting antibody epitopes to support epitope binning and epitope/function relationship analysis.
Our platform integrates machine learning along with structural biology, protein engineering, and translational immunology. At the center of our ability to harness the power of machine learning is our team’s deep expertise in biologics discovery and development, and our commitment to integrating different disciplines. Together, our platform and our people bring the capabilities needed to drive innovation for immunology drug discovery.
This links to an external website.