Project onboarding and resources
+
+
+Table of Contents
+ ++General resources and how-to’s for logistics and our technologies. +
+
+
+1. Server (moby)
+
+
+
+
+1.1. Install and setup tailscale
+
+
++Download and install tailscale. This will give you direct access to the server. +
+
+
+1.1.1. Macos installation from download
+
+
+
+
+/Applications/Tailscale.app/Contents/MacOS/Tailscale up --authkey tskey-abcdef1432341818 ++
+You will receive a machine key that you can use to enter the network. +
+
+
+1.1.2. Installed using brew/apt/…
+
+
+
+
+tailscale up --authkey tskey-abcdef1432341818 ++
+
+1.2. Select packages
+ +
+
+1.3. Add user to Nix Home manager
+
+
++This section ensures that the home manager system finds your files and generates your user. +
+-
+
- There are a couple of places left to edit to get your own user.
+
-
+
machine/moby/default.nix
+
+
+
+1.4. Day-to-day access
+
+
++You can access it in multiple ways. +
+
+
+1.4.1. Jupyter notebook
+
+
+
+This method requires screen both in remote and local computers.
+
+
+1.4.2. Step 1: Open Jupyter notebook on Moby
+
+
+
+
+ssh $USER@moby +screen +cd project/folder +# envrc should start by itself here. If this project has not been configured that way you can do: +# nix develop . --impure +poetry run jupyter notebook +# Copy the url including the token +# Press <Ctrl-A> -> <Ctrl-D> to detatch your screen session ++
+
+1.4.3. Step 2: Tunnel the port to your local computer
+
+
++Then open a tunnel to your jupyter notebook instance from the local computer. +
+
+
+
+screen -d -S tunnel -m 'ssh -L 8888:localhost:8888 $USER@moby -N'
+
++On your local computer, go yo the URL copied on step 1. +
+ ++If moby is restarted you must repeat step 1 (we can try to automate it too). +If your local computer is restarted you need to repeat step 2. +
+
+
+1.4.4. VS Code
+
+
++Tested on versions before 1.88. +
+ +-
+
- Install Remote Development Extension + +
+
+1.4.5. Emacs
+
+
++I use emacs for most things, so it is already installed in moby. +
+-
+
- Set up your packages of interest (doom emacs makes things easier). Make sure to include the direnv module. +
+2.Run +
+
+
+ssh $user@moby -t 'emacsclient -t' + ++
-
+
- You may need to fiddle a bit to get the proper environment working, this is work in progress. +
+
+2. Papers
+
+
++This may help you get more context +
+
+
+2.1. Morphological profiling
+
+
+-
+
- What is cell painting +
- The main dataset we will work on (Chandrasekaran et al. 2023). +
- Review on ML on morphological profiles (Chandrasekaran et al. 2021). +
+
+2.2. Statistics
+
+
+-
+
- How do we calculate reproducibility in Cell Painting experiments? (Kalinin et al. 2024) +
+
+2.3. Intepretable Deep Learning
+
+
+
+
+2.3.1. Counterfactuals
+
+
+-
+
- The basis of our plan. A new preprint will be released soon. (Eckstein et al. 2021) +
+
+2.3.2. Generative modelling
+
+
+-
+
- Recent work in the interface of morphological profiling and Generative Deep Learning (Lamiable et al. 2023) +
+
+3. Selected important events
+
+
+-
+
- Alán’s presentation with Janelia folks for a potential collaboration on Counterfactuals (See Counterfactuals). +
- ~ Mock Hackathon alongside CytoData to iron-out the issues and details necessary before the actual hackathon. +
- Hackathon organised by Alán, as part of SBI2-CytoData. +
+
+4. Bibliography
+
+
+
+
+Chandrasekaran, Srinivas Niranj, Jeanelle Ackerman, Eric Alix, D. Michael Ando, John Arevalo, Melissa Bennion, Nicolas Boisseau, et al. 2023. “JUMP Cell Painting Dataset: Morphological Impact of 136,000 Chemical and Genetic Perturbations.” bioRxiv. https://doi.org/10.1101/2023.03.23.534023.
+ Chandrasekaran, Srinivas Niranj, Hugo Ceulemans, Justin D. Boyd, and Anne E. Carpenter. 2021. “Image-Based Profiling for Drug Discovery: Due for a Machine-Learning Upgrade?” Nature Reviews Drug Discovery 20 (2): 145–59. https://doi.org/10.1038/s41573-020-00117-w.
+ Eckstein, Nils, Alexander S. Bates, Gregory S. X. E. Jefferis, and Jan Funke. 2021. “Discriminative Attribution from Counterfactuals.” arXiv. https://doi.org/10.48550/arXiv.2109.13412.
+ Kalinin, Alexandr A., John Arevalo, Loan Vulliard, Erik Serrano, Hillary Tsang, Michael Bornholdt, Bartek Rajwa, Anne E. Carpenter, Gregory P. Way, and Shantanu Singh. 2024. “A Versatile Information Retrieval Framework for Evaluating Profile Strength and Similarity.” Biorxiv, April, 2024.04.01.587631. https://doi.org/10.1101/2024.04.01.587631.
+ Lamiable, Alexis, Tiphaine Champetier, Francesco Leonardi, Ethan Cohen, Peter Sommer, David Hardy, Nicolas Argy, et al. 2023. “Revealing Invisible Cell Phenotypes with Conditional Generative Modeling.” Nature Communications 14 (1): 6386. https://doi.org/10.1038/s41467-023-42124-6.
+