L3 Bioinformatics
L3 Bioinformatics (L3B) is a pioneering company founded in 2014 as a spin-off from the University of Hong Kong’s Bioinformatics Algorithm Research Laboratory. It specializes in developing scalable, AI-driven solutions for genomic data analysis, bridging research and clinical practice. Their portfolio includes advanced tools like the Clair-series for variant calling, AutoPM3 for genomic interpretation, and secure data platforms such as Helicube and database.bio. L3B collaborates with top medical institutions to build integrated, intelligent systems that enhance data analysis speed, safety, and actionability, shaping the future of AI-powered genomics and digital health.
Industries
Nr. of Employees
small (1-50)
L3 Bioinformatics
Rm 2114, 21/F Hong Kong Plaza, 188 Connaught Road West, Hong Kong
Products
Deep-learning variant calling software
Neural-network based variant caller designed to detect SNPs and indels from short- and long-read sequencing, optimized for performance in high-error-rate data.
De novo metagenomic assembler
High-performance assembler for large and complex metagenomic datasets capable of single-node assembly without pre-partitioning, focusing on time- and resource-efficiency.
NGS analysis hardware-software platform
Integrated cluster management and workflow execution platform providing job scheduling, horizontal scalability, tool extensibility, and system monitoring for NGS analysis.
Genomics annotation and interpretation platform
Platform aggregating dozens of databases and many annotation types to support panel/exome/genome variant analysis with interactive review, smart filters, and report generation for clinical and research use.
Genomic variant web application
Secure web application for visualizing, prioritizing and annotating genetic variants using a configurable heuristic decision tree and links to supporting evidence.
Alignment visualization web application
Multi-track web browser for aligning and visualizing BAM/SAM pileups, BED regions, primers and reference sequences with support for multiple workstations linking to shared file hubs.
Deep-learning variant calling software
Neural-network based variant caller designed to detect SNPs and indels from short- and long-read sequencing, optimized for performance in high-error-rate data.
De novo metagenomic assembler
High-performance assembler for large and complex metagenomic datasets capable of single-node assembly without pre-partitioning, focusing on time- and resource-efficiency.
NGS analysis hardware-software platform
Integrated cluster management and workflow execution platform providing job scheduling, horizontal scalability, tool extensibility, and system monitoring for NGS analysis.
Genomics annotation and interpretation platform
Platform aggregating dozens of databases and many annotation types to support panel/exome/genome variant analysis with interactive review, smart filters, and report generation for clinical and research use.
Genomic variant web application
Secure web application for visualizing, prioritizing and annotating genetic variants using a configurable heuristic decision tree and links to supporting evidence.
Alignment visualization web application
Multi-track web browser for aligning and visualizing BAM/SAM pileups, BED regions, primers and reference sequences with support for multiple workstations linking to shared file hubs.
Services
Deployment and configuration of browser-based secure workspaces for sensitive clinical and genomic data with access controls, approval workflows, and audit logging.
Design, implement and integrate automated NGS workflows (alignment, variant calling, custom pipelines) with monitoring, visual pipeline editing, and recovery features.
Architecture and operational services for multi-site enterprise storage, hot/cold storage tiering, backup, and high-availability data farm operations.
Implementation of LLM-based pipelines to extract and structure literature evidence for variant classification and to accelerate manual interpretation workflows.
Deployment and configuration of browser-based secure workspaces for sensitive clinical and genomic data with access controls, approval workflows, and audit logging.
Design, implement and integrate automated NGS workflows (alignment, variant calling, custom pipelines) with monitoring, visual pipeline editing, and recovery features.
Architecture and operational services for multi-site enterprise storage, hot/cold storage tiering, backup, and high-availability data farm operations.
Implementation of LLM-based pipelines to extract and structure literature evidence for variant classification and to accelerate manual interpretation workflows.
Expertise Areas
- Clinical informatics and patient data integration
- Genomic data analysis and interpretation
- Bioinformatics algorithm development
- AI/ML for healthcare (deep learning, reinforcement learning, LLMs)
Key Technologies
- Deep learning for variant calling
- Reinforcement learning
- Large language models with retrieval-augmented generation
- Next-generation sequencing (NGS) pipelines