Researchers on LatchBio can now leverage nineteen AI protein tools in their browser, run 70% faster multi-omics pipelines, and achieve one-click access to NVIDIA accelerated computing.

LatchBio Simplifies Access to GPU-Powered Multi-Omics Tools With NVIDIA

Kyle Giffin | kyle@latch.bio
Kenny Workman | kenny@latch.bio

LatchBio is building data and pipeline infrastructure to help biotechnology teams access and analyze data at scale.

This press release features multimedia. View the full release here: https://www.bus

inesswire.com/news/home/20250206604264/en/

(Graphic: Business Wire)

The company is announcing today the release of nineteen accessible AI protein engineering tools, 70% faster multi-omics pipelines, and one-click access to NVIDIA accelerated computing.

Protein Engineering tools

The 2024 Nobel Prize in Chemistry was awarded for the development of groundbreaking new machine learning methods to understand proteins. To help researchers easily access these methods, LatchBio

is releasing nineteen optimized and publicly available protein engineering AI workflows, including AlphaFold2, RF-Diffusion and ProteinMPNN.

Each workflow can be run by a researcher in their Latch account, with a delightful user interface and managed compute – there is no cloud or IT setup required.

Early adopters, including Molecular Partners, are now screening tens of thousands of proteins on LatchBio to achieve faster pre-clinical results.

“The team achieved significant breakthroughs, demonstrating the success of our computational approach in designing molecules. Our aim is to build on this success and overcome limitations observed for other targets.” – Director of Oncology at Molecular Partners

Accelerated pipelines

Developers need access to fast and powerful pipelines to support the routine use of sequencing-based experiments that produce hundreds of gigabytes or terabytes of data.

Latch is partnering with multi-omics providers like AtlasXomics to re-rewrite their NGS pipelines with GPUs, achieving to up to 70% faster runtimes than existing methods. This is leading to performance breakthroughs that unlock new commercial applications.

“By rewriting the scATAC peak-calling algorithm to leverage NVIDIA GPUs, Latch was able to reduce our peak-calling time 14x, leading to significant reductions in total compute time and cost. We plan to continue working with Latch to drive down the cost of each step of our processing pipelines.” – James McGann, Bioinformatics Engineer at AtlasXomics.

One-click access to GPUs

Accelerated hardware can drive significant time and cost improvements in bioinformatics. But not all scientists have ready access to user-friendly bioinformatics infrastructure with GPUs.

LatchBio is also announcing the launch of Pods, a flexible cloud portal that helps researchers and scientists write, run, and share notebook code effortlessly on GPU hardware.

Built for machine learning, multi-omics, and scientific computing, Pods offer computational teams a seamless method for working together in reproducible environments on the cloud.

In one click, users can access pods with NVIDIA A10, L40S and A100 GPUs, allowing them to flexibly run GPU-intensive tasks in R, Python, or other languages in a secure environment.

“Access to GPUs is essential for our machine learning and deep learning work. We love that pods provide a user-friendly code development environment with access to GPU resources we need.” – Asmamaw (Oz) Wassie, ML & Bioinformatics Engineer at Glyphic Biotechnologies

Future Opportunities

This collaboration sets the stage for more accelerated infrastructure and pipelines, including faster single-cell workflows, multi-omics tools, and AI applications.

Kyle Giffin, founder & COO of LatchBio, commented, “New computational methods will facilitate many breakthroughs in biology the next 10-20 years. By focusing on ease of use, we’re ensuring scientists everywhere may access and benefit from these tools.”

To learn more about how LatchBio is accelerating bioinformatics with NVIDIA GPUs, visit https://latch.bio/.

View source version on businesswire.com: https://www.businesswire.com/news/home/20250206604264/en/