Using Spatial-Cell-ID computational resources through IFB cloud

A step by step guide

For analyzing single-cell RNASeq and Spatial Transcriptomics data, Spatial-Cell-ID users can exploit CBMPsmn dedicated resources through the IFB Cloud Biosphere.

Below are the instructions on how to proceed.

Join the Spatial-Cell-ID Group on IFB cloud

  1. Go to https://biosphere.france-bioinformatique.fr/

  2. Click on Sign in at the top right.

  3. Click Sign in again.

  4. Click on Accept conditions.

  5. Select your own institution.

  6. Log in with your institution’s username and password.

  7. Fill in the information: Last Name, First name, city, and postal code. Leave the other information as default and accept.

  8. You will be directed to a new page. Click on the small person icon at the top right, then select Groups.

  9. Click on Join a group in the tabs at the top left.

  10. Search for the group Spatial-Cell-ID and click on the + button to apply.

  11. That’s it! Now you have to wait for your application to be validated.

Once your application will be validated, you can deploy a Virtual Machine from Biosphere.

How to deploy a pre-configured Virtual Machine (VM) for single-cell RNA-Seq and Spatial Transcriptomics data analysis in Python

  1. Click on RAINBio at the top left.

  2. Select the appliance Spatial-Cell-ID-Jupyter.

  3. Click RUN at the top right and then select ADVANCED CONFIGURATION AND START.

  4. Choose a name for your VM, select Spatial-Cell-ID as the Group to use, meso-psmn-cirrus as the Cloud, and choose the appropriate Cloud flavor for your analysis

Note: For most single-cell RNA-Seq data analyses, the standard.4c16g configuration should be sufficient. However, for Spatial Transcriptomics, standard.16c64g is recommended. These are general suggestions; resource allocation should be determined on a case-by-case basis. Please do not select a configuration that exceeds your actual requirements, as this avoids wasting computational resources and energy.

Please be mindful of the environment!

  1. Go to myVM at the top right.

  2. Wait for the appliance to start. At some point, you will see that Access will have two options: https and params. Click on params to view the user ID and password, then click on https to connect to the virtual machine using the user ID and password you just obtained.

  3. That’s it! Your virtual machine is ready.

  4. That’s it! Your virtual machine is ready. To open a new Jupyter notebook with the pre-configured computational environment for single-cell or spatial transcriptomics analysis, click the blue “+” button in the top-left corner and select “spatialdata” under the “Notebook” section. You will find installed most of the libraries you will need, e.g. scanpy, squidpy and spatialdata.

Similarly, you can open a terminal within the same environment by clicking the “+” button and selecting the “Terminal” icon. Once the terminal is open, ensure the environment is active by running conda activate spatialdata to access your bioinformatics tools via the command line. Using the terminal allows you to potentially install new tools into the environment using either conda or pip, ensuring your workspace remains flexible for specific analysis needs.

Beyond Spatial-Cell-ID-Jupyter Virtual Machine.

While the Spatial-Cell-ID-Jupyter appliance is specifically tailored for spatial transcriptomics and comes with pre-loaded libraries, the IFB Biosphere catalogue offers a wide variety of other pre-configured Virtual Machines (VMs) for different research needs.

If you require a more flexible or generic workspace, you might consider deploying one of these standard environments:

  • Ubuntu 24.04: Ideal for users who wish to work exclusively through a command-line interface in a standard Linux environment.

  • Ubuntu 24.04 Desktop : A full Ubuntu 24.04 installation featuring a Graphical User Interface (GUI). This requires the X2Go client to access the desktop remotely. For setup instructions, please refer to the IFB documentation on X2Go.

  • RStudio Server: Optimized for data analysis and visualization using the R programming language.

  • Jupyter: A standard environment for Python-based data science workflows.

Important Note: Unlike the specialized Spatial-Cell-ID-Jupyter appliance, these general-purpose VMs (Ubuntu, RStudio, and Jupyter) provide only the base environment.

Unlike the specialized Spatial-Cell-ID appliance, these general-purpose VMs (Ubuntu, RStudio, and Jupyter) provide only the base environment. In these cases, you are responsible for installing all necessary tools and configuring your own computational environment. To do that, you can use the terminal to install any specific packages you need via conda or pip.

In principle, you are allowed to use any VM listed in the catalogue. However, please be aware that most appliances are configured for very specific biological analyses. For general tasks where you want to build your own pipeline from scratch, the standard Ubuntu, RStudio, or Jupyter images are your best starting points.

Please, refer to the IFB documentation for further information.

Important: Data Persistence and Storage Safety

Managing Your Deployment

It is crucial to understand the difference between closing your session and terminating your VM:

  • Safe to Close: You can shut down your local computer or close the web browser tab where the virtual machine is running without losing any work. You can resume your session at any time via the MyVM section.

  • Data Loss Warning: If you click on Terminating deployments, the Virtual Machine is permanently deleted. Any data, scripts, or configurations saved within the VM (outside of persistent storage) will be lost forever.

  • Data Transfer: Uploading and downloading files to/from your VM is intuitive, but please feel free to ask for help if you are unsure how to secure your files.

Permanent Storage (~/ifbdata/spatial_cell_id)

To allow you to move between different VMs or keep data after a termination, a persistent storage area is available. This directory is shared across every VM you open.

  • Path: ~/ifbdata/spatial_cell_id/

  • Usage: New users should create their own dedicated subfolder here (e.g., ~/ifbdata/spatial_cell_id/your_name/).

CRITICAL WARNINGS!

  1. No Privacy or Protection: The spatial_cell_id folder does not have a backup or privacy system. There is no “undo” for deleted files—once a file is removed, and it cannot be recovered.

  2. Shared Access: This is a shared space. Every user can technically access, modify, or delete data belonging to others. Please exercise extreme caution and strictly interact only with your own folder.

  3. Analysis Space Only (Not an Archive): This directory is not a long-term storage system. It is provided solely as a workspace to maintain continuity while you are actively performing your analysis or switching VMs.

  4. Mandatory Clean-up: Once your analysis is complete, you must move your data to a permanent external storage and delete it from this folder. The Spatial-Cell-ID team is not responsible for long-term data hosting; it is your responsibility to secure your results and scripts elsewhere.

Environmental Responsibility and Resource Management

We are committed to the sustainable and efficient use of computational resources. Please be mindful that running Virtual Machines consumes significant energy and infrastructure capacity.

  • Power Down: We kindly ask you to shut down or terminate your VMs whenever they are not in active use.

  • Avoid Idle Sessions: Do not leave VMs running for extended periods (such as days or weeks) unless you have a specific, long-running calculation that requires it.

Because we provide permanent storage for your work, you can safely terminate your VM at the end of your workday and simply reconnect your data to a new instance when you are ready to resume.

Resource Quotas and Impact

Our access to the IFB cloud is based on a fixed allocation of compute hours. Every wasted hour directly affects the entire community:

  1. Shared Credits: We have a “deal” for a specific total of hours with the IFB. Once these hours are exhausted, the entire group will lose access to IFB resources until a new allocation is granted.

  2. Core Multiplier: Credits are consumed based on the size of your VM. If you deploy a VM with 16 cores, you are consuming 16 compute hours for every 1 hour of real time the machine remains active.

Please be responsible: only request the resources (CPU/RAM) you actually need for your analysis, and terminate your deployment as soon as your work is done.

Contacts

If you have any questions, inquiries, or need assistance with the environments and storage systems described in this tutorial, please do not hesitate to contact Sergio Sarnataro, Head of the Data Analysis Hub at Spatial-Cell-ID (Email: sergio.sarnataro@ens-lyon.fr).