Skip to content
R&D

AI in drug discovery: Key trends shaping therapeutics in 2025

4 min read

The pharmaceutical industry is undergoing a profound transformation, with AI-driven drug discovery revolutionizing how new therapeutics are developed. Machine learning (ML), including deep learning and generative AI, is accelerating drug candidate identification, optimizing molecular structures, and reducing costs. Key applications, such as AI for target identification, virtual screening, and predictive modeling, are streamlining drug development and increasing success rates. 

AI is reshaping peptide-based drug discovery by enabling rapid design, activity prediction, and optimization of novel therapeutics. These advances bypass traditional trial-and-error methods, expediting the path from concept to candidate selection. 

AI drug development techniques

AI-driven peptide discovery

AI-driven approaches, such as deep learning models and predictive screening, are transforming peptide discovery by enabling the design and selection of potent drug candidates at unprecedented speed. 

Traditional peptide drug discovery has been hindered by limited native peptide ligands and labor-intensive optimization processes. Now, machine learning models can design peptides from scratch, optimizing their properties with high accuracy. 

Gubra’s AI-powered drug discovery approach

Gubra is leveraging AI-driven methodologies to accelerate peptide-based drug discovery. Central to this is de novo peptide design, where entirely new sequences are generated to fit a desired target. By integrating AlphaFold for structure prediction and generative models like proteinMPNN—which proposes amino acid sequences that are compatible with a given 3D backbone—, Gubra’s platform designs and optimizes novel peptides with high receptor potency, and beneficial drug-properties. 

Machine Learning (ML)-guided peptide optimization

Beyond AI-driven de novo peptide design, Gubra also leverages machine learning for peptide optimization through its drug discovery platform, streaMLine. This platform combines high-throughput data generation with advanced AI models to guide the selection of the most promising drug candidates. By integrating experimental data with predictive modeling, streaMLine enables informed decision-making in the drug discovery process. In a parallelized setup, the platform simultaneously optimizes for potency, selectivity, and stability, thus accelerating timelines and the success rate of new drug candidates. 

streaMLine platform graphic

The streaMLine platform

How streaMLine is transforming peptide drug development

streaMLine represents a paradigm shift in peptide drug discovery, integrating AI into optimization of peptides. By leveraging machine learning, the platform systematically refines peptide candidates, ensuring they meet stringent drug-like criteria such as efficacy, specificity and long half-life. 

One of the key applications of streaMLine was demonstrated in the development of novel GLP-1 receptor (GLP-1R) agonist based on a secretin backbone. By leveraging AI-guided design, the platform: 

  • Enhanced receptor selectivity: AI-driven substitutions improved GLP-1R affinity while abolishing off-target effects.
  • Optimized stability: Modifications reduced peptide aggregation and improved solubility. 
  • Achieved long-acting efficacy: In vivo studies demonstrated potent weight-loss effects in diet-induced obese (DIO) mice, and a PK profile compatible with once-weekly dosing. 

This AI-powered approach underscores the impact of machine learning in drug discovery, accelerating the transition from in silico modeling to functional peptide drug candidates. 

AI in drug discovery

Deep learning models like AlphaFold and RFdiffusion are enabling the design of drug candidates for previously ‘undruggable’ disease targets

The future of AI in drug discovery at Gubra

As AI technologies continue to evolve, their role in drug discovery will expand to encompass a broader range of therapeutic modalities. Gubra remains committed to integrating advanced machine learning methodologies into its drug discovery pipeline, ensuring that future therapeutics are optimized for efficacy, safety, and manufacturability. 

Gubra is at the forefront of AI-driven peptide therapeutics, pioneering next-generation drug discovery through machine learning innovation. Stay ahead of the curve and partner with Gubra to shape the future of your therapeutic pipeline with AI-powered drug development. Explore our latest research on AI-powered drug development and learn how Gubra is shaping the future of peptide drugs. 

Machine-learning-guided peptide drug discovery:
Development of GLP-1 receptor agonists with improved drug properties

Read the full paper here

Pipeline for development of acylated peptide based
CGRP receptor antagonist with extended half-life for migraine treatment

Read the full paper here

Categories
R&D
Cite this article
"AI in drug discovery: Key trends shaping therapeutics in 2025" in Gubra, Mar 19, 2025, https://www.gubra.dk/blog/ai-in-drug-discovery-key-trends-shaping-therapeutics-in-2025/.
Copy Citation
Related posts
No entries found.

For further information

Contact us

Gubra

Hørsholm Kongevej 11B
2970 Hørsholm
Denmark

[email protected]
+45 3152 ­2650

Back To Top