Skip to content
October 11, 2021

Reducing the Complexity of Single-Cell Genomic Analysis

How GPU acceleration and Code Ocean Compute Capsules Make Complex Work Flows Faster and Easier

Single cell analysis of DNA and RNA is an important method that allows researchers to characterize cell populations with increased resolution.

Single cell analysis of DNA and RNA is an important method that allows researchers to characterize cell populations with increased resolution.

“Combining single-cell data with complementary datasets can lead to exciting discoveries and new insights; AI, specifically machine learning (ML), is fast becoming the gold standard used to perform important tasks like dimensionality reduction and unsupervised clustering.”     Emily So, Solutions Scientist at Code Ocean

However, the advantages of working on the single cell level come at a price: growing data sets require computing power that computational researchers might not have at their disposal and the complexities of setting up stable analysis pipelines can limit adoption.

To overcome these challenges three critical components are needed:

  • faster processing using GPU acceleration
  • flexible environments that allow for easy deployment of models and
  • overall reduction of the complexity of the process so computational researchers can more readily implement them.

In our webinar Accelerating Single-Cell Genomics Discovery with AI, NVIDIA, and Code Ocean Emily So, Solutions Scientist at Code Ocean, discusses how combining AI/ML algorithms with NVIDIA-powered GPU acceleration in the flexible and user-friendly Code Ocean Compute Capsules helps computational researchers scale and speed up single-cell analysis.
To illustrate Emily presents two Compute Capsules:

  • RAPIDS GPU Compute Capsule (starting at 18:45 mins)
  • ATACworks Compute Capsule (starting at 27:30 min)

Accelerating single cell genomic analysis

The RAPIDS GPU Capsule contains example notebooks that demonstrate how to use RAPIDS for GPU-accelerated analysis of single-cell sequencing data. The presentation:

  • introduces the different components of Compute Capsule – metadata, environment, code, data and results,
  • discusses how to set-up and customize them, e.g. by adding compute resources, installing dependencies, libraries and packages and selecting the IDE needed to run the code,
  • shows how to use a Jupyter Notebook for cell type clustering and biomarker discovery to create and visualize the results,
  • demonstrates how Capsules can be shared, specifically, how the App Panel enables sharing with colleagues who have limited coding experience and prefer a no-contact, one-click way of rerunning the analysis.

ATACWorks Denoising data and calling peaks

  • The ATACworks Capsule contains the code, environment and data to denoise and identify accessible chromatin regions from low-coverage or low-quality ATAC-seq data . The demonstration shows how users can:
  • run the code in a Jupyter Notebook inside a Compute Capsule, including importing data and adding required packages to generate the results
  • review and analyze the results of denoising of the data and calling of the peaks that indicate chromatin accessibility.

For any questions about GPU acceleration, the RAPIDS and ATACWorks Capsules or about Compute Capsules in general, please book a demo.

Tag(s): Bioinformatics

Projects We're Working On:

View All Posts