Virtual screening with the GOLD Cluster Computing package

Our 2021.1 release allows high throughput docking with GOLD to be run with HPC clusters, giving you flexibility to perform virtual screens with cloud or cluster compute power.

Previous updates allowed ultra-large docking, or virtual screening, with GOLD to be run on cloud resources. This presents a flexible and low-cost option to engage additional computational power in an ad-hoc manner. For example, see our whitepaper where this approach was used in a 126 million ligand screen. However, we understand that many sites may prefer to use their own high performance computing (HPC) cluster. 

This update allows you to apply the accurate, high-quality docking of GOLD at scale, with flexibility in deployment.

virtual screening with GOLDVirtual screening with GOLD has been deployed in various studies, including this work looking for actives against COVID-19 Mpro - read the full story here.

 

GOLD virtual screening requirements

  • System requirements
    • The cluster will need to have SingularitySlurm and Python 3 (3.6 or later) installed. No extra Python modules are necessary.
    • Please note that large virtual screens will require a lot of disk space for the input ligands. For instance, 100 million ligands downloaded from ZINC as gzipped SDF files takes 60 GB of disk space (the equivalent MOL2 files would be larger). The batching that is required for GOLD docking on the cluster (e.g. 50,000 batches of 2,000 ligands) would add an additional 100 GB (we use a lower level of compression for speed). Please make sure you have enough disk space for your job size. For large scale screens, we recommend use of the 'MIN_OUT' option to reduce the amount of output from GOLD. 
  • Installation or downloads needed:
    • Download the GOLD Cluster Computing package (see link below) and extract it in a suitable location on the cluster.
    • A CSD installation is not required on the system as GOLD is included inside the Singularity image (a SIF file) that is supplied with the package, but a suitable licence is required.
  • Licence requirements
    • You will need a CSD-Discovery licence to use the GOLD Cluster Computing package.

 

The GOLD Cluster and Cloud products will be available for you to access at the following location https://www.ccdc.cam.ac.uk/support-and-resources/downloads/ under CSD-Discovery->GOLD as 'GOLD Cluster Computing' and 'GOLD Cloud Computing' links.

 

Learn more or start virtual screening with GOLD

If you would like to get started with this, do get in touch our Support Team to help you get going.

Learn more: