October 2025 Summaries

Single-cell sequencing is a cutting-edge research technique that isolates and sequences the DNA or RNA of individual cells, providing a high-resolution view of cellular heterogeneity that bulk sequencing misses. It is crucial for understanding complex biological processes, such as disease mechanisms in cancer and rare diseases, by revealing cellular variations and enabling precise diagnostics and therapeutic targets. There are three main types of single-cell sequencing: genome, transcriptome, and DNA epigenomic sequencing, with multi-omics approaches combining these techniques for greater insights. Despite its advantages, including detecting rare cell types and detailed lineage relationships, single-cell sequencing faces challenges like high costs, technical complexity, and data sparsity. The cost for single-cell sequencing can range from $1,700 to $3,800 per sample, influenced by factors like sequencing depth and platform choice. Databases such as the Human Cell Atlas and GEO provide platforms for storing and sharing single-cell sequencing data, facilitating research and discovery. TileDB offers solutions to optimize computational and storage costs, enhancing the efficiency and scalability of single-cell analysis.

TileDB Carrara addresses the challenge of making diverse data AI-ready by introducing omnimodal intelligence, enabling the governance, structuring, and analysis of various data types within a single system. Traditionally, data is siloed, governance is inconsistent, and complex data like genomics is difficult to process, hindering AI utilization. Carrara treats all data and code as first-class modalities, allowing teams to query and analyze assets on a unified platform powered by TileDB's multidimensional array engine, which provides the speed and scalability needed for intensive research. This approach is particularly transformative for life sciences, as it facilitates cohort-scale analytics, reproducible workflows, and infrastructure cost reduction by integrating complex data and AI-ready pipelines. Carrara's four pillars—organize, structure, collaborate, and analyze—enhance data management and collaboration, promoting FAIR data principles. The FAIR Readiness Scorecard helps organizations evaluate their data maturity to further advance AI infrastructure.

TileDB offers a comprehensive platform for single-cell RNA sequencing analysis, designed to efficiently handle the complexities of transcriptomics data on a large scale. TileDB Carrara, built on multi-dimensional arrays, powers the specialized TileDB SOMA system, enabling seamless integration with popular analysis tools like Seurat, Bioconductor, and Scanpy, while being optimized for cloud storage solutions such as Amazon S3 and Google Cloud Storage. This platform supports scalable data management and analysis for projects with over 100 million cells, as demonstrated by its use in the Chan Zuckerberg CellxGene Atlas. TileDB Carrara facilitates efficient data unification, query performance, and cost-effective storage, while also enabling advanced features such as vector search for cell similarity and annotation, support for spatial transcriptomics, and high-performance table operations. The platform is designed to streamline research workflows by transitioning from file storage to active analysis, providing secure collaboration environments, and supporting AI/ML model training with optimized data workflows.

TileDB Carrara is a platform designed to streamline bioinformatics workflows by integrating data management, computational processes, and interactive analysis into a single, governed environment. It addresses the challenges of managing diverse data types inherent in multiomics analysis, such as raw sequencing files, reference genomes, and specialized databases, by treating them as first-class citizens within a unified catalog. The platform supports secure collaboration through Teamspaces, which manage permissions and security without duplicating data, and offers native support for Nextflow workflows, allowing seamless transitions from raw data to queryable databases. Carrara facilitates debugging and monitoring with task graph visualizations and supports re-entrant retries, enhancing workflow efficiency. Upon workflow completion, results are immediately accessible for interactive analysis, and the platform's notebook environment allows for direct query of data with integrated terminal access. Carrara also excels in high-performance table operations using Apache DataFusion, enabling SQL queries and transformations on complex genomic data. This comprehensive approach positions Carrara as both a daily driver for analysis and a Trusted Research Environment, promoting secure data sharing and reproducibility in life sciences research.

TileDB, Inc., Kepler AI, and Tahoe Therapeutics have announced a partnership to create a platform that allows researchers to run AI agent-based queries on massive single-cell datasets, marking a significant advancement in computational biology. This collaboration integrates TileDB's multimodal database technology, Kepler AI's agent systems, and Tahoe Therapeutics' comprehensive single-cell datasets, resulting in the Tahoe-100M Atlas, a giga-scale single-cell perturbation atlas that is now openly accessible. The platform, which stores over 100 million transcriptomic profiles from 50 cancer cell lines, allows AI agents to work directly with full-resolution single-cell data, providing researchers with real-time analysis capabilities and enabling the discovery of new biomarkers and therapeutic pathways. The platform, available at tahoebio.ai/tahoedive, offers a scalable cloud-native solution to facilitate drug discovery and genomics research, while lowering the barrier for machine learning teams to conduct experiments and train larger models. The collaboration is expected to significantly accelerate research timelines in drug susceptibility, resistance analysis, and genotype-specific transcriptional response characterization.

TileDB has announced the controlled availability of TileDB Carrara, an omnimodal data intelligence platform designed to address the challenges of managing diverse data types in a unified environment. Carrara revolutionizes data management by treating all data types, from structured tables to complex non-tabular data such as genomics and imaging, as structured and governed modalities. This approach overcomes the limitations of traditional data handling methods, which often result in siloed and inefficient data management. The platform's four pillars—organization, structuring, collaboration, and analysis—facilitate seamless data discovery, secure sharing, and efficient analysis across various industries, with significant implications for the life sciences and healthcare sectors. With its future-ready architecture, Carrara is adaptable to emerging data types, enabling organizations to innovate and scale without significant infrastructure changes. TileDB has forged partnerships with major technology providers like Databricks and Snowflake and is supported by cloud giants AWS and Microsoft Azure, with global integrators such as Cognizant and Accenture enhancing its digital transformation initiatives.