Reading
This article introduces SQUALL, a 555-million-parameter multimodal foundation model that integrates histopathological images and spatial transcriptomics data to predict molecular expression from pathological images, providing a new tool for cancer research and clinical diagnosis.
Section 01
SQUALL is a 555-million-parameter multimodal foundation model that integrates histopathological images and spatial transcriptomics data to predict molecular expression from pathological images, providing a new tool for cancer research and clinical diagnosis. Developed by the OswaldZhang team, its open-source code repository was released on GitHub on June 1, 2026.
Section 02
In the fields of cancer research and precision medicine, histopathology and spatial transcriptomics are core diagnostic technologies, but they have long been in silos. Traditional methods process image and genetic data separately, lacking a unified model, which limits the ability to infer molecular features and predict prognosis from pathological images.
Section 03
SQUALL adopts a phased self-supervised training strategy: 1. Image feature learning; 2. Cross-modal alignment; 3. Task fine-tuning. Based on the histMol corpus (33 tissue types, 12 platforms, 1.76 billion spatial spots), its core capabilities include large-scale virtual biomarker analysis (covering 15,757 genes) and identification of prognosis-related spatial microenvironments (e.g., TLS maturation patterns).
Section 04
In breast cancer research, it identified molecular trajectories related to tumor invasion; in ovarian cancer, it recognized recurrence-related immune-excluded microenvironments; its performance in prognosis prediction tasks outperforms traditional models, and it shows balanced performance in predicting platinum chemotherapy resistance, providing support for clinical decision-making.
Section 05
The open-source code repository includes modules for preprocessing, pre-training, fine-tuning, and inference, based on the PyTorch framework. It provides tutorials and examples, supporting the generation of virtual gene expression profiles from pathological slides and spatial microenvironment clustering analysis.
Section 06
It breaks through the boundaries of multimodal spatial representation learning, promotes low-cost molecular typing, retrospective cohort studies, real-time diagnostic assistance, and drug response prediction, lowering the technical threshold for precision medicine.
Section 07
Current limitations include insufficient coverage of training data (bias towards rare cancers) and limitations in spatial transcriptomics resolution; future plans include integrating higher-resolution data, expanding cancer types, and combining clinical information to build a comprehensive model.
Section 08
SQUALL integrates histopathology and spatial molecular data, advances the boundaries of computational pathology technology, provides a practical tool for precision medicine and cancer research, and is expected to play an important role in clinical practice in the future.
Keep going with more reads from the same topic.
Nornir MCP Server is an enterprise-level server based on the Model Context Protocol (MCP). It seamlessly integrates large language models (such as Claude) with the Nornir network automation framework, supporting natural language orchestration for multi-vendor network devices (Cisco, Arista, Juniper, etc.), and providing production-grade features like a dual-engine architecture (NAPALM + Netmiko), intelligent filtering, and a secure sandbox.
Bibliothèque Française LLM is a structured indexing and annotation project for French public domain literature designed specifically for large language models (LLMs). It integrates multiple authoritative sources such as DraCor, Common Corpus, and Wikisource, providing metadata indexing categorized by genre, author, and era, as well as in-depth annotations for dramatic texts (including characters, lines, stage directions, etc.). Its aim is to enable LLMs to efficiently read and understand classic French literary works.
Splinter is a minimalist, high-performance key-value (KV) and vector storage system enabling zero-latency inter-process communication via shared memory and atomic operations. With only 766 lines of core code, it supports millions of operations per second and 768-dimensional vector storage, offering a new architectural approach for local LLM inference and data-intensive applications.
Folkering OS is the world's first AI-native bare-metal operating system, entirely written in Rust no_std without relying on Linux, POSIX, or libc. It can generate commands from scratch, compile them into WASM, and run them in 10 seconds, achieving true self-evolution.