diff --git a/content/blog/2025-conference-summary.md b/content/blog/2025-conference-summary.md
new file mode 100644
index 0000000..3740c73
--- /dev/null
+++ b/content/blog/2025-conference-summary.md
@@ -0,0 +1,182 @@
++++
+title = "Highlights from the scverse conference 2025"
+date = 2025-11-20T00:00:05-08:00
+description = "A recap from the second scverse conference at Stanford University"
+author = "Giovanni Palla"
+draft = false
++++
+
+
+
+ Attendees gather for a group photo at Stanford University
+
+
+The 2nd scverse conference at Stanford University, Palo Alto, California, on November 17-19, 2025 brought together developers and scientists from academia and industry to discuss the latest advances in computational tools and experimental techniques for single cell and spatial multi omics.
+
+
+
+ The conference in numbers
+
+
+The conference in numbers:
+
+- 4 keynote speakers
+- 15 invited talks
+- 3 sponsor talks
+- 90 posters
+- 250 attendees in person
+- 100 attendees on zoom
+- 3 parallel workshop sessions at the Arc Institute (Palo Alto), Biohub (Redwood City) and Tahoe Therapeutics (South San Francisco)
+
+The conference was sold out one month in advance!
+
+
+
+ Our generous sponsors and travel award recipients
+
+
+Thanks to our generous sponsors, the scverse conference was also able to provide:
+
+- Travel awards to 9 awardees from 5 different countries.
+- Conference fee waivers to 16 attendees from 7 different countries.
+
+### Keynote Speakers
+
+
+
+ John Marioni, Genentech
+
+
+**John Marioni**
+*Senior VP and Head of Computation at Genentech Research and Early Development*
+
+In his talk *"Adventures at the interface between computational biology, AI and drug discovery"* he discussed the intersection of computational biology, artificial intelligence, and pharmaceutical drug discovery.
+
+**Panos Roussos**
+*Professor at Icahn School of Medicine at Mount Sinai*
+
+Panos Roussos presented advances from large-scale single-cell and spatial transcriptomic studies in neurodegenerative and neuropsychiatric diseases in his talk *"Deciphering Interindividual Variability in Brain Transcriptomes Through Single-Cell Analysis"*.
+
+**Erika Alden DeBenedictis**
+*Co-founder of The Align Foundation & CEO of Pioneer Labs*
+
+In her talk *"How do we make the next PDB? Introducing Align's Data Platform"* she described [Align Foundation](https://alignbio.org/)'s mission and their end-to-end data generation process: community roadmapping → protocol standardization → reproducible data collection → model benchmarking.
+
+**Elham Azizi**
+*Associate Professor of Cancer Data Research and Biomedical Engineering at Columbia University*
+
+In her talk *"Machine Learning for Decoding Tumor Microenvironment Dynamics"* she discussed the latest work from her lab on how machine learning frameworks in spatial transcriptomics, histology imaging, and single-cell sequencing are providing new insights in how the tumor microenvironment (TME) shapes cancer progression and therapy response.
+
+### Short Talks Summary
+
+#### Day 1: Perturbation Profiling, Functional & Regulatory Genomics
+
+The first day centered on advances in perturbation screening and regulatory genomics. Several talks explored innovations in Perturb-seq technology, including clonal lineage tracking to reveal lineage-specific responses to genetic perturbations, Bayesian frameworks for scalable differential expression analysis and optimal experimental design in CRISPR screens, and minute-scale automated scRNA-seq for real-time perturbation profiling. The afternoon shifted toward regulatory genomics and machine learning applications, featuring zero-shot prediction of regulatory element activity across tissues, interpretable deep learning for mapping genetic effects at cellular resolution in population-scale data, and attention-based methods for predicting CAR T-cell therapy outcomes from single-cell data.
+
+#### Day 2: Alternative Splicing, Spatial Biology & Tools
+
+The second day broadened the scope to RNA splicing, spatial technologies, and ecosystem tooling. A pair of talks addressed the often-overlooked dimension of alternative splicing — one revealing isoform landscapes in spatial transcriptomics, the other tackling robust integration of sparse splicing and expression data. Large-scale perturbation datasets for training dose-dependent foundation models and methods for decoding sequence determinants of gene expression were also presented. The afternoon highlighted tool development across modalities: an updated peak caller adapted for single-cell genomics, graph-based frameworks for spatial and subcellular proteomics, surface-based cell interaction analysis in 3D spatial omics, and interactive visualization tools bridging the scverse ecosystem with modern web-based interfaces.
+
+
+
+### Agentic Workflows in Bioinformatics Panel Discussion
+
+At the end of the first day, the conference hosted a panel discussion on agentic workflows in bioinformatics, providing an opportunity to the audience to hear from leaders in the field on the state of the art of agentic systems in biological data analysis. The panel consisted of:
+
+- **Loic Royer**: Research Director and Group Leader at Biohub San Francisco
+- **Vega Shah**: NVIDIA Ventures Portfolio Manager
+- **Kexin Huang**: PhD student at Stanford University and leading the Biomni Project
+- **Jonah Cool**: Head of Life Sciences Partnerships at Anthropic
+
+The discussion revolved around diverse topics, from technical discussions on [Model Context Protocols (MCPs)](https://modelcontextprotocol.io/) and Context Engineering to a forward-looking perspective on the impact of agentic systems in the work of bioinformaticians.
+
+### Workshop Sessions
+
+
+
+
+ Participants of the Biohub Workshop
+
+
+
+ Interactive session at Biohub
+
+
+
+#### Processing and Modeling Large-Scale Single-Cell Perturbation Data with Machine Learning @ Arc Institute
+
+The Arc Institute workshop provided an end-to-end journey through Perturb-seq workflows. The morning covered practical considerations for designing and running Perturb-seq experiments, followed by sessions on data processing including a new evaluation suite for single-cell perturbation models, differential expression analysis approaches, and strategies for working with large AnnData stores. The afternoon shifted to machine learning, featuring an AI-agent curated resource of uniformly processed single-cell data spanning over 500 million cells, perturbation effect modeling with [scvi-tools](https://scvi-tools.org/), modular tokenization frameworks for single-cell foundation models, and methods for predicting perturbation effects across diverse cellular contexts. The day concluded with a panel discussion evaluating progress toward a virtual cell and planning for the next [Virtual Cell Challenge](https://virtualcellchallenge.org/).
+
+#### Multimodal And Spatial Single Cell Data: From Data APIs to AI-driven Biological Discovery @ Biohub
+
+The Biohub workshop balanced infrastructure, acceleration, and community-driven tool development. Sessions introduced the CELLxGENE foundation and community landscape, followed by a deep dive into GPU-accelerated single-cell analysis in collaboration with NVIDIA. Community showcase sessions highlighted diverse contributions: handling 3D spatial omics data, accessible whole slide image analysis integrated with the scverse ecosystem, and a trio of tools spanning graph-based single-cell analysis, proteomics imaging, and spatial visualization. Additional sessions covered the [Virtual Cell Platform](https://virtualcellmodels.cziscience.com/) model development ecosystem, benchmarking approaches, and AI workspaces for collaborative research.
+
+#### Disease Biology and Therapeutics in the Age of Frontier Datasets @ Tahoe Therapeutics
+
+The Tahoe workshop emphasized computational scalability and the emerging role of AI agents in single-cell analysis. Morning sessions covered GPU-accelerated analysis at scale with rapids-singlecell and scalable data loading pipelines for deep learning on large single-cell datasets. A notable thread focused on AI agents and LLM-based tools, including building custom MCP servers, an agent for drug response signature interpretation, and agents for large-scale biomedical dataset mining. The afternoon featured Tahoe-x1, a perturbation-trained single-cell foundation model scaled to 3 billion parameters, alongside a practical session on enrichment analysis with decoupler. A panel discussion closed the day.
+
+### Poster Session
+
+
+
+
+
+
+
+
+During the two days of the conference, a total of 88 posters were presented, where participants had the chance to engage in deeper discussions about their work across the scverse ecosystem.
+
+### Sponsor Talks
+
+The conference also featured 3 talks from our diamond sponsors:
+
+- *Towards a virtual cell signaling model rooted in human development* by Samantha Dale Strasser (Somite Therapeutics)
+- *Powering the Virtual Cell: Frontier Datasets for AI-Driven Biology* by Johnny Yu (Tahoe Therapeutics)
+- *Solving complex challenges in human health at the intersection of AI and biology* by Angela Oliveira Pisco (Biohub)
+
+Each sponsor provided an outlook of the latest technologies in perturbation screens, multimodal data integration, and frontier AI and biology.
+
+### Testimonials
+
+
+"scverse is the conference to be at to stay up-to-date on all the current happenings in the single-cell community"
+
+
+
+
+
+
"Exceptional conference! The scverse conference achieved true single-human resolution with every attendee perfectly clustered, well-annotated, and deeply engaged."
+
+
+
+
"As a user rather than a developer of computational tools, I discovered several resources, methods, and workflows that I did not know existed, which will directly improve the way I approach my analyses. What stood out most was how open and collaborative the community is, making advanced tools accessible even for those who focus primarily on applying, rather than building, computational methods."
+
+
+
+
"A good mix of industry and academic presence, the conference provided an insight into the bleeding edge of methods but also the ways in which they are being used."
+
+
+
+
"The scverse community has accelerated science by crafting solutions to core, underappreciated problems. I admire the team's commitment to openness and transparency, and how that promotes scientific integrity in their growing community."
+
+
+
+
"The workshops were amazing, the science was top-tier, and the smaller setting made it so easy to connect with people. The scverse organizers were incredibly hardworking and thoughtful—they made the entire experience smooth, generous, and deliciously well-catered."
+
+
+
+
+### Conclusion
+
+This was the second time we organized the scverse conference — you can read about [the first edition in Munich](/blog/2024-conference-summary/). We're looking forward to seeing you at future scverse events! In the meantime, you can engage with the community on [Zulip](https://scverse.zulipchat.com/) and at our community meetings. Look out for announcements on our social media channels ([X](https://x.com/scaboranova), [Bluesky](https://bsky.app/profile/scverse.bsky.social), [LinkedIn](https://www.linkedin.com/company/scverse/)).
+
+*Photo credits: Vivekanandan Ramalingam (Vivek), Research Scientist at Stanford Genetics.*
diff --git a/static/img/blog/conference2025/conference2025-day2.jpg b/static/img/blog/conference2025/conference2025-day2.jpg
new file mode 100644
index 0000000..fa45995
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-day2.jpg differ
diff --git a/static/img/blog/conference2025/conference2025-keynote.jpg b/static/img/blog/conference2025/conference2025-keynote.jpg
new file mode 100644
index 0000000..ae46614
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-keynote.jpg differ
diff --git a/static/img/blog/conference2025/conference2025-numbers.png b/static/img/blog/conference2025/conference2025-numbers.png
new file mode 100644
index 0000000..57b7513
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-numbers.png differ
diff --git a/static/img/blog/conference2025/conference2025-poster1.jpg b/static/img/blog/conference2025/conference2025-poster1.jpg
new file mode 100644
index 0000000..68f0f4d
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-poster1.jpg differ
diff --git a/static/img/blog/conference2025/conference2025-poster2.jpg b/static/img/blog/conference2025/conference2025-poster2.jpg
new file mode 100644
index 0000000..554d4a8
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-poster2.jpg differ
diff --git a/static/img/blog/conference2025/conference2025-poster3.jpg b/static/img/blog/conference2025/conference2025-poster3.jpg
new file mode 100644
index 0000000..89efc8b
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-poster3.jpg differ
diff --git a/static/img/blog/conference2025/conference2025-poster4.jpg b/static/img/blog/conference2025/conference2025-poster4.jpg
new file mode 100644
index 0000000..178c540
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-poster4.jpg differ
diff --git a/static/img/blog/conference2025/conference2025-shorttalk.jpg b/static/img/blog/conference2025/conference2025-shorttalk.jpg
new file mode 100644
index 0000000..445cfd7
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-shorttalk.jpg differ
diff --git a/static/img/blog/conference2025/conference2025-sponsors.png b/static/img/blog/conference2025/conference2025-sponsors.png
new file mode 100644
index 0000000..69936af
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-sponsors.png differ
diff --git a/static/img/blog/conference2025/conference2025-workshop1.jpg b/static/img/blog/conference2025/conference2025-workshop1.jpg
new file mode 100644
index 0000000..0dbd5d3
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-workshop1.jpg differ
diff --git a/static/img/blog/conference2025/conference2025-workshop2.jpg b/static/img/blog/conference2025/conference2025-workshop2.jpg
new file mode 100644
index 0000000..1bc8311
Binary files /dev/null and b/static/img/blog/conference2025/conference2025-workshop2.jpg differ
diff --git a/static/img/blog/conference2025/conference2025.jpg b/static/img/blog/conference2025/conference2025.jpg
new file mode 100644
index 0000000..16a0dd8
Binary files /dev/null and b/static/img/blog/conference2025/conference2025.jpg differ