Come explore how AI and machine learning are reshaping scientific discovery and healthcare innovation. This meetup will spotlight two transformative applications of ML in the life sciences—one in clinical diagnostics and pathology, and the other in foundational model building for cutting-edge omics research.
Whether you're a seasoned data scientist, a healthcare professional, or simply passionate about the future of biotech, this is the perfect place to connect, learn, and be inspired.
***
Featured Talks: AI Across the Spectrum of Healthcare and Discovery
1. Using AI for Generating Pathology Reports
Speaker: Aaron Hardin, Director, Clinical Biomarker Discovery, Guardant Health
Pathology is the cornerstone of diagnostic medicine, but it remains complex and time-consuming. In this talk, Aaron Hardin will discuss how Guardant Health is leveraging AI to streamline the generation of pathology reports—enhancing accuracy, speed, and consistency to support better clinical decision-making in oncology and beyond.
2. Training foundational models for tissue and cfDNA methylation Data
Speaker: Radhakrishna (RK) Bettadapura, Vice President, Research Informatics, Strand Life Sciences
Abstract:
DNA methylation is a fundamental epigenetic mechanism that regulates gene expression and significantly impacts cellular differentiation, genomic stability and disease development. We present foundational models for both tissue and cell-free DNA methylation and illustrate their utility by fine-tuning them for 2 applications i) early cancer detection in cfDNA methylation panel sequencing of 220 samples with various cancer types ) ii) Chronological age prediction in ~1300 samples in AltumAge, a multi-tissue DNA methylation dataset. Our foundational models capture complex relationships and higher-order interactions between CpG sites making use of positional, sequence context as well as methylation fraction information.
The tissue-based model was trained on methylation sequencing data from 75k samples (totaling 1.1 billion CpG sites) across different methylation array platforms, including Illumina 27k, 450k, EPIC, EPIC+, and EPICv2 arrays. It employs an encoder-only Performer transformer architecture to efficiently handle 10k randomly sampled CpG sites. The pre-training phase involves an imputation task, where the model predicts beta values for masked CpG sites, focusing on learning accurate representations of methylation patterns. Notably, the learned embeddings cluster by CpG island annotation, despite the model never receiving this information as input—indicating that it captures underlying genomic structure in an unsupervised manner. Results demonstrate reasonable performance in beta-value imputation, achieving a Pearson correlation coefficient of ~0.89. Fine-tuning can be performed to adapt the pre-trained model for downstream applications such as cancer classification and age prediction.
***
Agenda
- 5:00 – 6:00 PM: Check-in & Networking
- 6:00 – 7:30 PM: Welcome, Presentations & Q&A
- 7:30 – 8:30 PM: Continued Networking
Light refreshments will be provided.
***
Venue
Guardant Health
3100 Hanover St, Palo Alto, CA 94304
A huge thanks to Guardant Health for hosting and sponsoring this event!
***
Who Should Attend?
- Bioinformaticians
- Computational biologists
- Data scientists in biopharma and healthcare
- Healthcare professionals and technologists interested in AI/ML applications
If you're excited about how machine learning is changing the future of healthcare and research, you won’t want to miss this.
***
Want to Get Involved?
- Have an AI/ML use case or product to showcase?
- Interested in sponsoring, hosting, or helping organize future meetups?
- Just love geeking out about AI in healthcare?
We'd love to hear from you! Reach out at admin@eppicglobal.org.
***
About EPPIC Global
EPPIC Global is a cross-disciplinary network bringing together healthcare professionals, scientists, technologists, and entrepreneurs. From early discovery to clinical application, our community is united by a shared mission: driving innovation to create better healthcare outcomes.
We organize events, mentorship programs, and our flagship EPPICon conference—helping bridge the worlds of technology and medicine through connection, collaboration, and visionary thinking.
