Overview

EPI2ME Desktop from Oxford Nanopore Technologies is a user-friendly standalone application that allows researchers and bioinformaticians to perform genomic data analysis. While it can be used on Windows and MacOS desktops, some workflows are hardware resource intensive. In addition, basecalling workflow requires a GPU and is only supported on Linux.

This guide will walk you through the steps to install EPI2ME Desktop on NVIDIA GPU Amazon EC2 instance with Ubuntu Linux to run basecalling and other workflows. It also covers how you can scale your EC2 vCPU count and memory size, and switch between GPU and non GPU instance.

Other options

If you need a managed cloud based Virtual Desktop Infrastructure (VDI) solution, you can consider Amazon WorkSpaces. Do ensure that your workflow minimum system requirements can be met (up to 8 vCPUs and 32 GB memory for Ubuntu WorkSpaces Personal).

For fully managed bioinformatics workflows at scale, consider AWS HealthOmics.

System Requirements

EPI2ME Basecalling workflow has the following requirements

• Recommended: 64 vCPUs and 256 GB memory
• Minimum: 8 vCPUs and 64 GB memory
• NVIDIA GPU with at least 8 GB of vRAM

The above requires 16xlarge (64 vCPUs/256GB RAM) and 4xlarge (16 vCPUs/64 GB RAM) accelerated computing G or P instance size respectively, e.g. G7e, G6, Gr6, G6e, G5 and G4dn

EPI2ME Somatic Variation workflow has the following requirements

• Recommended: 64 vCPUs and 256 GB memory
• Minimum: 16 vCPUs and 48 GB memory

The above requires 16xlarge (64 vCPUs/256 GB RAM) and 4xlarge (16 vCPUs/64GB RAM) general purpose M instance size respectively, e.g. M8a, M8i, M8i-flex, M7a and M7i, M7i-flex

GPU Instance Types

Login to your console, go to EC2 console, Instance Types to verify which GPU instance types are available for your Region

Note the instance type and Availability Zone where the GPU instance is available

Service Quota

Go to Service Quota console for your Region, and verify applied quota value for the following

• Running On-Demand G and VT instances
• Running On-Demand Standard (A, C, D, H, I, M, R, T, Z) instances

You will need a value of 16 and 64 to meet the minimum and recommended requirements respectively.

Select the quota from the list and choose Request increase at account level if otherwise. Do not proceed until your applied quota value is equal or higher than minimum required vCPU count

Prepare EC2 instance

Launch a new EC2 instance with the following

• Ubuntu Linux 24.04/22.04
• x86_64 NVIDIA GPU EC2 instance
• 100 GB or more EBS volume size
• Security group that allows inbound TCP and UDP port 8443 from your IP for DCV access
• EC2 IAM role with AmazonS3ReadOnlyAccess managed policy attached for DCV license validation

The following needs to be installed

• NVIDIA GPU driver
• Docker Engine with NVIDIA Container Toolkit
• Desktop environment and manager
• Amazon DCV server
  • Configure virtual sessions and multi-user.target to allow EC2 instance type change between GPU and non GPU families

Below are a few options to install NVIDIA driver, Docker with NVIDIA Container Toolkit and Amazon DCV server.

For Option 1 and 2, configure DCV virtual session and multi-user.target run level. You will need to create a virtual session manually for user login. Option 3 provides managed virtual sessions.

Option 1: Install NVIDIA software and DCV manually

Launch a NVIDIA GPU instance running Ubuntu Linux, and use the following community articles to install NVIDIA software, graphical desktop and DCV server

• Install NVIDIA GPU driver, CUDA Toolkit, NVIDIA Container Toolkit on Amazon EC2 instances running Ubuntu Linux
  • Install AWS Command Line Interface (AWS CLI) (optional)
• Install GUI (graphical desktop) on Amazon EC2 instances running Ubuntu Linux
  • Configure DCV virtual session and multi-user.target run level
  • Create virtual session

Option 2: Install DCV on DLAMI

If available in your Region, you can install graphical desktop and DCV server on AWS Deep Learning AMI

• Deep Learning graphical desktop on Ubuntu Linux with AWS Deep Learning AMI (DLAMI)
  • Configure DCV virtual session and multi-user.target run level
  • Create virtual session

Option 3: Use CloudFormation

You can use CloudFormation template from this Github repository to provision EC2 instance with NVIDIA driver, Docker with NVIDIA Container Toolkit, Amazon DCV server with managed virtual session, and AWS CLI.

From Ubuntu-NICE-DCV.yaml page, select Download raw file.

Go to CloudFormation console, and create a new stack with the downloaded template file.

Use the following settings

• Key Pair: select your EC2 key pair
• OS version and processor architecture: Ubuntu 24.04 (x86_64)
• Instance type (x86_64 or arm64): specify a x64_64 NVIDIA GPU instance type available in your Region
• DCV session type and GPU driver:
  • G6f, Gr6f instance: virtual-with-NVIDIA_GRID_Driver
  • Others: either virtual-with-NVIDIA_repo_Driver or virtual-with-Ubuntu_repo_Driver

• Volume size (GiB): at least 100. Enter higher value if you intend to run multiple workflows

• Install Docker Engine : Yes (default)
• Select a VPC and subnet in selected VPC with outbound IPv4 internet connectivity. Ensure your desired instance type is available in the selected subnet (AZ)

Refer to Github page for more details.

Connect to EC2 instance

Connect to EC2 instance as ubuntu

Installation

Install Java

sudo apt install -y openjdk-17-jdk

Verify that the correct Java version is installed

java -version

Output should be similar to below

openjdk version "17.0.17" 2025-10-21
OpenJDK Runtime Environment (build 17.0.17+10-Ubuntu-122.04)
OpenJDK 64-Bit Server VM (build 17.0.17+10-Ubuntu-122.04, mixed mode, sharing)

Install Nextflow

Install Nextflow

cd
curl -s https://get.nextflow.io | bash

Output should be similiar to below

      N E X T F L O W
      version 25.10.2 build 10555
      created 28-11-2025 19:24 UTC
      cite doi:10.1038/nbt.3820
      http://nextflow.io


Nextflow installation completed. Please note:
- the executable file `nextflow` has been created in the folder: /home/ubuntu
- you may complete the installation by moving it to a directory in your $PATH

Configure Nextflow

We will move Nextflow to an executable path ~/.local/bin

mkdir -p ~/.local/bin/
mv nextflow ~/.local/bin/
sed -i '$aexport PATH=$PATH:$HOME/.local/bin/' ~/.bashrc
. ~/.bashrc

Updating Nextflow

To update Nextflow,

nextflow self-update

Output should be similar to below

ubuntu@ip-172-31-29-57:/tmp$ nextflow self-update

      N E X T F L O W
      version 25.10.3 build 10983
      created 22-01-2026 15:34 UTC 
      cite doi:10.1038/nbt.3820
      http://nextflow.io


Nextflow installation completed. Please note:
- the executable file `nextflow` has been created in the folder: /home/ubuntu/.local/bin
- you may complete the installation by moving it to a directory in your $PATH

Install EPI2ME Desktop

Install EPI2ME Desktop

mkdir -p /tmp/epi2me
cd /tmp/epi2me
wget https://ont-exd-int-s3-euwst1-epi2me-labs.s3.amazonaws.com/installers/epi2me_5.3.0_amd64.deb
sudo apt install -y ./epi2me*.deb

Using EPI2ME Desktop

Connect to EC2 instance Amazon DCV session. Locate and launch EPI2ME.

Refer to EPI2ME Desktop documentation for usage guidance. Below image is EPI2ME Desktop on g5.16xlarge instance running Ubuntu Linux 24.04

Scale EC2 Instance

You can change the EC2 instance type and size to scale vCPU count and memory size, or to switch between accelerated computing GPU and general purpose non-GPU instance.

1. Stop the EC2 instance
2. After instance state has changed to Stopped, choose Actions, Instance settings, Change instance type
3. On the Change instance type page, enter an appropriate instance type
4. Click Change instance type to apply the change
5. Start your EC2 instance

Registry of Open Data on AWS and Oxford Nanopore Technologies Open Data

You can use AWS Command Line Interface (AWS CLI) to download datasets from Registry of Open Data on AWS, which also includes Oxford Nanopore Technologies Open Data project (ont-open-data) reference sequencing data. Refer to Oxford Nanopore Open Data Tutorials for more information