Use high-throughput structural modeling to accurately predict exposed liabilities for thousands of sequences at once
Before the advent of next-generation sequencing (NGS), drug developers searching for new therapeutic antibodies could only rely on low-throughput, time-consuming technologies that required years to develop just a handful of candidates. With NGS technology, researchers have gained access to unprecedented levels of sequencing depth that allow them to identify millions of receptor sequences in a single experiment.
However, this presents a new problem: it is no longer possible – or economically feasible – to express and functionally test millions of individual antibody candidates. So how can you efficiently evaluate the entire candidate pool, and come to a shortlist of diverse antibodies with the highest potential and the lowest risk?
The importance of fast and accurate calculation of sequence liabilities and developability risks
Compared to Sanger sequencing, bulk and single-cell NGS technologies generate orders of magnitude larger volumes of data, thereby significantly expanding the available candidate pool. In theory, the ability to pick from a large and diverse pool of antibodies should go hand in hand with a higher likelihood of finding antibodies with the most favorable (clinical) properties. However, due to the sheer volume of generated data, it’s impossible to test all candidates to assess their developability. This means that many candidates can fail late in the development process, when significant amounts of time and resources have already been spent. To increase candidate success, it is crucial to couple a diverse candidate pool (as offered by NGS) with a fast and accurate way to assess sequence liabilities and the developability of candidates early in the discovery phase. This enables researchers to:
- Reduce the risk of picking candidates unfit for production and clinical use.
- Decrease time spent on antibody engineering.
- Obtain more first-time-right, high-quality candidates.
The dilemma of liability predictions
Historically, the assessment of antibody developability has been a trade-off between throughput and accuracy. Screening for specific sequence motifs associated with liabilities in linear sequences scales well but doesn’t consider the location of motifs in the context of the 3D structure of the antibody. Crucially, exposed liabilities that are found on the surface of the antibody have the highest likelihood to interfere with the downstream development and production of an antibody. By only assessing the linear sequence, motifs that are located in the inner portions of the antibody can falsely flag an antibody for elimination.
However, while offering highly accurate predictions, structural modeling of antibodies comes with a big challenge due to its need for significant computational resources.
IGX-Annotate offers accurate liability annotations at scale
To combine a best-in-class user experience with great accuracy and throughput, ENPICOM developed a specialized App for the IGX Platform called IGX-Annotate. It provides high-throughput structural modeling of antibodies to accurately identify and annotate exposed liabilities. This new App is powered by the Structural Antibody Prediction Platform (SAbPred): a validated, peer-reviewed, and globally recognized toolbox for accurate and efficient structural analysis of antibodies that was developed by researchers at the University of Oxford.
In addition to accurate liability predictions, IGX-Annotate provides an assessment of the developability of candidates by comparing the characteristics of your antibodies to those of clinically approved antibodies.
- Leverage a validated tool for liability identification. Use a validated structural modeling toolbox to improve the liability annotation accuracy.
- Annotate structural liabilities for thousands of sequences. Run structural modeling with SabPred for many molecules in parallel to scale the liability analysis to your needs.
- Directly compute developability profiles. Configurable penalty scores for all predicted liabilities and developability assessments allow you to effectively identify and prioritize the antibody candidates that fit your research strategy.
- Overlay liability scores. Computed scores can directly be used to guide the selection of your top candidates, e.g., by overlaying them in visualizations like the phylogenetic tree viewer, or to identify clusters of interest.
From thousands of sequences to that needle-in-a-haystack antibody, the IGX Platform allows you to diversify your candidate pool and easily integrate accurate developability predictions into your workflow to simply select better antibodies. Be sure to read our release notes and download our latest white paper to learn more about the latest IGX Platform features.
Discover how the IGX Platform can help you accelerate and improve your antibody discovery. Talk to our experts today.