To empower clinicians and researchers by transforming proteomic and clinical data into actionable insights through a reproducible machine learning pipeline, advancing precision care for Multiple Sclerosis patients.
Harnessing transcriptomics and ML to predict MS progression.
Renewable and sustainable energy sources for a better & the greener future.
Get Started
Predict. Personalize. Prevent. Progress.
5
4
3
2
1
4
3
2
1
0
5
4
6
3
2
1
5
4
6
3
2
1
0+
0
0
1
0
0
0
0
1
0
0
Patients Analyzed
FusionMS: a solution for clinicians, neurologists, and biomedical researchers.
Your future starts here—with tailored data insights advancing precision MS care.
Get started
What We Do
IntegraMS bridges the gap between proteomics research and clinical practice by transforming complex molecular and patient data from consented research cohorts into interpretable, predictive models for Multiple Sclerosis (MS). Our cloud-based platform, FusionMS, enables clinicians and researchers to upload, analyze, and visualize proteomic and clinical data through an intuitive, web-based interface. FusionMS applies reproducible machine learning to model disability severity (DSS) and to explore proteins and immune reactivity that may be associated with disease progression. Rather than providing clinical recommendations, our platform is designed to support discovery and exploratory translational research, helping teams turn high-dimensional data into insights that can inform future studies.
.svg)
Why We Do It
Multiple Sclerosis is a complex neurological disorder: its causes are multifactorial, symptoms vary widely, and disease trajectories differ substantially between patients. While advances in molecular profiling and biomarker detection are revealing more about the underlying biology of MS, these data are rarely analyzed alongside longitudinal clinical outcomes in a unified manner. Researchers face fragmented datasets spread across experimental methods and time horizons, while clinicians often lack the tools to link these research findings to meaningful clinical context. This disconnect slows translational progress and limits the broader impact of promising discoveries.
IntegraMS exists to close this gap. By uniting proteomic and clinical data under a transparent and interpretable machine-learning framework, we aim to accelerate MS research and support clinician-scientists working at the intersection of discovery and patient care. Our long term vision is building a future for more personalized and predictive approaches to MS care.
How We Do It
IntegraMS combines a secure cloud-based data pipeline, an interactive analytics interface, and a reproducible machine learning engine to deliver multimodal MS research at scale. FusionMS, built with React and Tailwind, allows users to upload structured proteomic and clinical datasets in CSV format, launch standardized analyses, and explore results through interactive dashboards. Behind the scenes, a Node.js backend manages authentication, data ingestion, validation, and storage using AWS S3 and Firebase, providing a scalable and reliable foundation. The pipeline performs peptide-level filtering and feature engineering on over 500,000 peptide-antibody measurements and integrates these features with clinical metadata to model disability status and disease trajectories. A hierarchical synthetic data generator built with the Synthetic Data Vault (SDV) framework further enhances testing and scalability by creating realistic synthetic cohorts that improve robustness and demographic balance, while maintaining a focus on research validity rather than clinical decision making. Results are served through REST APIs and visualized in real time, enabling researchers to ask more integrated questions about immune activity and MS progression using data that are increasingly common across immunology and neurological disease research.
FusionMS: A Clear, Reproducible Workflow From Data to Insight
Streamlined Work Process That Drives Results Efficient, Transparent, and Tailored for Your Success
Contact Us
01
02
03
04
01
Data Acquisition & Integration
FusionMS begins by securely collecting proteomic (phage display), neurofilament light chain (NfL), and clinical metadata from appropriately consented research cohorts. Data sources may include public repositories and controlled-access datasets such as those referenced by UCSF Neurology and the Department of Veterans Affairs. All data are ingested through a cloud-based pipeline built on AWS S3 and Firebase, ensuring compliance, reproducibility, and security.
02
Exploratory Data Analysis (EDA)
The platform performs structured exploratory analysis, including data cleaning, normalization, and feature inspection across 500,000+ protein measurements per patient. FusionMS handles missing values, identifies outliers, and visualizes molecular and clinical trends to highlight patterns of immune reactivity and identify features for downstream analysis.
03
Machine Learning & Prediction
FusionMS applies reproducible machine-learning workflows, including regression based and neural network models, to model disability severity (DSS) and explore relationships between proteomic signals and clinical outcomes. To support robustness and scalability, the platform incorporates Synthetic Data Vault (SDV) based cohort generation, enabling bias assessment and validation in a research setting.
04
Interactive Visualization & Results
Analytical results are delivered via REST APIs in an interactive React/Tailwind dashboard, where users can upload datasets, launch analyses, and explore outputs in dynamic charts. This interface enables researchers to examine the complex relationships between proteomic and clinical data in a transparent and repeatable manner, supporting discovery efforts and translational MS research.
Our Talented Team
Meet our UC Berkeley MIDS team of data scientists, engineers, and researchers dedicated to advancing precision care for Multiple Sclerosis through innovation in machine learning and biomedical data science.
Akshay Sharathchandra
Project Lead
Maria Jose Healey
Head of ML Engineering
Abhay Naik
Lead Application Architect
Coco Sun
Product Lead
Surabhi Gupta
Lead Systems Engineer
Here are some frequently asked questions
Find answers to the most common questions about how IntegraMS works and how it’s helping bring data-driven precision to Multiple Sclerosis care.
What is Multiple Sclerosis (MS)?
Multiple Sclerosis is a chronic autoimmune disorder where the body’s immune system attacks the protective coating of neurons, disrupting nerve signals between the brain and body. This leads to symptoms such as weakness, numbness, vision problems, and cognitive changes .
What is Proteomic, Omic, and Phage Data?
Omics refers to large-scale biological datasets, such as genomics or proteomics. Proteomic data specifically measures proteins and their activity in the body. Phage display data captures protein–antibody interactions using bacteriophage particles, helping researchers identify protein signatures that may indicate disease onset or progression .
Who Can Use IntegraMS?
IntegraMS is built for neurologists, clinical researchers, and data scientists. Clinicians can use it to support decisions on disease progression and subtype risk, while researchers can explore links between molecular data and patient outcomes .
What’s Next for IntegraMS?
Future development focuses on integrating additional multi-omics layers (genomic and transcriptomic), enhancing model interpretability, and scaling cloud-based deployment for real-world validation with clinical partners .
How Does IntegraMS Protect Patient Data?
All datasets are fully de-identified before processing. The platform employs encryption, secure access controls, and HIPAA/GDPR-compliant cloud infrastructure to ensure patient privacy