chelsie.li

What Are Neoantigens and How Does AI Enhance Their Discovery?

Neoantigens, arising from somatic mutations within tumor cells and absent in normal tissues, are unique markers crucial for cancer immunotherapy. They enable personalized medicine by allowing the immune system to effectively target and attack cancer cells. The discovery and application of neoantigens are revolutionizing cancer treatment, offering highly targeted therapies with reduced side effects compared to traditional methods ^[1].

The integration of artificial intelligence (AI) in neoantigen discovery has significantly improved the precision and efficiency of identifying viable neoantigens from complex array of tumor mutations. Advanced platforms leverage machine learning algorithms to unearth promising neoantigen candidates from multiomics data, incorporating diverse biological features crucial to cancer immunotherapy responses ^[3]. For instance, the platform ImmuneMirror, spearheaded by Chuwdhury and colleagues, showcases a robust method that utilizes experimentally validated immunogenic neopeptides to ensure high reliability and precision. This technology not only optimizes the identification process but also sharpens the selection of neoantigens with superior immunogenic potential, essential for crafting effective personalized immunotherapies ^[3].

As the field advances, AI's role in neoantigen research is evolving, integrating increasingly complex datasets and refining algorithms to navigate the vast diversity and intricacies of cancer genomics. These AI tools have transformative potential in cancer immunotherapy, promising more effective treatments in the future.

Beyond Discovery: Why Neoantigen Production Matters?

While discovering neoantigens is crucial, their production is equally vital, especially given the urgent timelines of cancer immunotherapy. Effective treatments often hinge on narrow therapeutic windows, necessitating rapid production and delivery of neoantigens. This production phase, encompassing synthesis, purification, and quality control, is where neoantigens are prepared for therapeutic use, making timing a critical factor.

'Intelligent Production': Streamlining Neoantigen Manufacturing Processes

A collaboration of AI tools, for example peptide production difficulty prediction (PepPre), which includes a specialized submodule for Neoantigen prediction (NeoPre 2.0), alongside mass-spectrum result identification (MsIde), HPLC retention time prediction (RetPre), purification efficacy prediction (PurPre), and peptide solubility prediction (SolPre), developed by bioinformatics scientists at GenScript can significantly enhance manufacturing processes. These innovations are continuously using extensive and reliable historical data, proactively addressing potential delays and optimize schedules for synthesis, purification, and quality control. As a result, cancer treatment times are reduced, ensuring swift and effective delivery of personalized therapies to patients.

The Benefits of Being 'Intelligent', 36% faster per item now.

At GenScript, AI models consider a spectrum of factors influencing production timelines in neoantigen production. These factors include the characteristics of peptide sequences, required quantity, purity specifications, and the synthesis challenges. Scientists integrate these variables into specialized algorithms tailored for specific tasks. For instance, peptide production difficulty prediction (PepPre) model incorporates cutting-edge techniques such as self-attention mechanisms and convolutional neural networks. This model demonstrates exceptional performance metrics with an overall accuracy of 91% and a precision of 64.9% ^[4]. This advanced approach significantly reduces production time shortening the cycle by an average of 36% per item ^[5]. This marks a significant achievement in an industry where every second counts.

Additionally, these AI models do more than just speed up the manufacturing of neoantigens, they also improve the quality and consistency of the peptides produced. Post-synthesis tools such as MsIde, RetPre, and PurPre refine production pipeline to ensure that peptides meet the highest standards of purity and quality. Meanwhile, SolPre supports both production team and customers by offering peptide solubility testing with a temporal accuracy of 80% ^[6], and it plays a crucial role in the design and evaluation of products even before finalization.

The success of AI-driven strategies highlights the potential of smart technologies in biomanufacturing, setting new industry standards and showcasing how GenScript is leading biotechnological innovations. AI tools support both the discovery and production of therapeutic neoantigens, hastening the availability of personalized cancer treatments and enhancing patient outcomes.

Curious About Our Tools or Interested in Testing Your Order with Our AI-Assisted Production Capabilities?

Contact us today at peptide@genscript.com. Discover how our cutting-edge solutions can benefit your specific needs.

Reference：

1. Xie, N., Shen, G., Gao, W., Huang, Z., Huang, C., & Fu, L. (2023). Neoantigens: promising targets for cancer therapy. Signal transduction and targeted therapy, 8(1), 9.

2. Cai, Y., Chen, R., Gao, S., Li, W., Liu, Y., Su, G., ... & Zhang, X. (2023). Artificial intelligence applied in neoantigen identification facilitates personalized cancer immunotherapy. Frontiers in Oncology, 12, 1054231.

3. Chuwdhury, G. S., Guo, Y., Chiang, C. L., Lam, K. O., Kam, N. W., Liu, Z., & Dai, W. (2024). ImmuneMirror: A machine learning-based integrative pipeline and web server for neoantigen prediction. Briefings in Bioinformatics, 25(2), bbae024.

Remarks:

4. Model Construction and Validation: The models were constructed using production data (beyond 0.4M items) generated before 2024. Their performance was rigorously tested with actual production orders handled by the GenScript Peptide Line during the first half of 2024. This approach ensures that our models are both current and directly applicable to real-world manufacturing scenarios.

5. Data Integrity and Calculation: All data used for model training and performance evaluation are meticulously calculated and managed by the order management team of GenScript Peptide Line, ensuring high data integrity and relevance.

6. Ongoing Model Refinement: Initially, our models were developed using a select subset of our comprehensive production data. We are actively collecting much more data, upgrading and optimizing these models to enhance their precision and efficiency further, ensuring that they evolve in line with technological advancements and production needs.