case study

Optimizing Pipeline Execution with High Performance Computing

Life Sciences Healthcare

Business Impacts

80% reduction in pipeline execution time

Improved performance and scalability

Reduced time and effort by leveraging terraform templates

Customer Key Facts

  • Country : US East
  • Industry : Healthcare

Problem Context

The client is a drug discovery company that wants to accelerate the development of allosteric drug therapies to identify drug regulatory hotspots.

The client sought to build their DNA Encoded Library (DEL) Pipeline on a cloud-native platform. The client also wanted to accelerate the process of building their protein libraries and reduce the management overhead by leveraging High Performance Computing on Google Cloud.

Challenges

  • Existing pipeline required more than 35 days to run tasks for 25 million molecules
  • Unable to run heavy computing workloads on the same HPC cluster
  • Long and complicated setup time for CompChem

Technologies Used

Compute Engine

Compute Engine

SLURM

SLURM

Pub/Sub

Pub/Sub

Cloud Storage

Cloud Storage

Cloud Functions

Cloud Functions

Firestore

Firestore

Cloud SQL

Cloud SQL

Building DNA-encoded Library (DEL) on a cloud-native platform

Solution

  • Quantiphi helped client develop a pilot environment and set up a SLURM-based HPC cluster that houses a DNA-encoded library pipeline in a cloud-native manner
  • Quantiphi also assisted the client by providing terraform scripts to automate their project setup

Results

  • Ease of management and operability
  • Significant acceleration in pipeline with the execution time reduced from 35 days to under 6 days for 25 million molecules
  • One-click deploy solution developed using Terraform, allowing the client to delete High Perfromance Computing clusters as required
  • Fault tolerant and highly available solution designed with the High Performance Computing cluster designed in a shared environment, allowing teams to utilize the cluster

Thank you for reaching out to us!

Our experts will be in touch with you shortly.

In the meantime, explore our insightful blogs and case studies.

Something went wrong!

Please try it again.

Share