AI-Driven Translational Research and Multi-Specific Antibody Discovery for Atopic Dermatitis

In immunology research, discovering effective therapies increasingly depends on connecting translational insights with scalable AI-driven analysis. This case study highlights how a biotechnology company used AI to improve both multispecific antibody design and the understanding of treatment non-response in atopic dermatitis. By combining target discovery with clinical insight analysis, the organization accelerated therapeutic research and uncovered new opportunities for precision immunotherapy.

Nextigent AI implemented two interconnected AI platforms to support antibody discovery and translational research for atopic dermatitis.

Healthcare
9 months combined
10+ specialists across AI, bioinformatics, and clinical analytics

high-potential co-targets identified for IL-18 therapies

consistently upregulated targets identified in Dupixent non-responders

↓

Research turnaround reduced from months to weeks

50%+

Hypothesis-to-validation cycle reduced by 50%+

Project Scope

The engagement combined two parallel research initiatives:

AI-assisted multispecific antibody discovery
AI-powered translational analysis of treatment non-responders

The scope included biomedical data integration, AI model development, feature selection, clustering analysis, interactive visualization, and deployment within the client’s secure research infrastructure.

Tech Stack

Python
PyTorch
TensorFlow
Vertex AI
Spotfire

Data sources

OpenTargets
Human Protein Atlas
STRING
GEO
ClinVar
TCGA
PubMed

The Client

The client is a biotechnology company specializing in immunotherapies and research on inflammatory diseases.Their goal was twofold:

improve efficacy of IL-18 monoclonal antibody therapies through multispecific design
better understand the molecular mechanisms behind non-response to Dupixent treatment in atopic dermatitis patients

Business Requirements

Identify synergistic co-targets for IL-18 inhibition
Discover biomarkers and therapeutic hypotheses for treatment non-responders
Reduce time required for translational analysis and target discovery
Integrate omics, serology, and clinical datasets into unified workflows
Support iterative and scalable AI-assisted research processes

The Problem

The client faced two major research bottlenecks.First, identifying biologically meaningful target combinations for multispecific antibodes required evaluating thousands of protein interactions across complex omics datasets.Second, existing analytical workflows could not efficiently explain why some atopic dermatitis patients failed to respond to Dupixent therapy.Traditional methods relied heavily on manual investigation, fragmented tools, and time-consuming collaboration between research teams.

The Challenges

Integrating highly heterogeneous translational and omics datasets
Detecting subtle biological correlations within high-dimensional data
Ensuring the interpretability of AI-generated insights for scientific teams
Supporting real-time hypothesis refinement and iterative simulations
Maintaining secure deployment within a regulated research environment

Problems Revealed

During implementation, it became clear that valuable biological signals existed across datasets but were not connected through a unified analytical framework.Target discovery, translational analysis, and clinical interpretation were operating in parallel rather than as part of a continuous research workflow. This limited the organization’s ability to systematically identify synergistic therapies and non-response mechanisms.

The Solution

Nextigent AI deployed two complementary platforms:

The Target Synergy Discovery AI Platform
The Clinical Insight Discovery AI Platform

The Target Synergy Discovery system analyzed multi-omics, pathway, and expression datasets to identify co-targets that enhance IL-18 inhibition in atopic dermatitis. Large language models supported literature mining and pathway analysis to uncover biologically relevant combinations.The Clinical Insight Discovery platform focused on translational analysis of Dupixent non-responders. AI-driven clustering and feature selection models analyzed serology, expression, and clinical response data to identify consistently upregulated targets and biomarker patterns.Interactive dashboards enabled scientists to explore correlations, visualize differential expression, and refine hypotheses in real time across both workflows.

Achievements

Identified 7 high-potential IL-18 co-targets
Discovered 5 consistently upregulated targets in Dupixent non-responders
Reduced translational analysis timelines from months to weeks
Reduced antibody discovery cycles by more than 50%
Improved collaboration between immunology, translational, and data science teams

The combined platforms transformed fragmented research activities into a structured, AI-assisted discovery ecosystem.

Future Plans

The client plans to extend both platforms into additional inflammatory and autoimmune disease programs.Future development includes predictive modeling for treatment response and expanded support for multispecific therapeutic design workflows.

Ready to see how enterprise-grade security  can accelerate your AI initiatives?

Schedule a Technical Security Briefing

Request Security Review