How to setup a mac for genomic analysis

Question

[nitinra]

Hello all,

I have been doing all my genomic analysis on a remote cluster so far and it works perfectly for me. However, there are times when I wish I could run it on my own computer (since I have to request permissions every time to install modules etc). I have a macbook pro and I am looking for a tutorial which will guide me to set my laptop to run these analysis. Specifically, I am interested in learning how to install base packages such as python, numpy, etc.

Thank you.

[ATpoint]

There is not much that is really Mac-specific, it is Unix after all. You must to install XCode which is Apple's development environment. For compilation of R packages it makes sense to download the latest gfortran and clang as described here https://cran.r-project.org/bin/macosx/tools/. I would recommend to get Homebrew installed (the macOS-tailored package manager) and install the coreutils (so the GNU coreutils that you are used to have on Linux) via brew install coreutils which includes basics like ls and sort which are superior to the annoying macOS in-builts. conda as suggested below makes sense. You can perfectly run two package managers in parallel. Homebrew is often faster and more convenient plus macOS-tailored, but conda can be more powerful when managing multiple environments. For co-existance of Homebrew and conda you should run echo 'auto_activate_base: false' >> ~/.condarc so conda is not activated by default to avoid conflicts with Homebrew. You can then always activate conda environments as needed, please refer to their docs. By the way, you can perfectly work on a cluster environment with conda as well, I do this for years and never asked anyone to install any application for me since I manage almost everything via conda, but I agree, a laptop can sometimes be convenient. That is basically what came to my head right away, feel free to comment if anything is unclear.

Answer 1

[GenoMax]

Use conda part from this tutorial: Creating workflows with snakemake and conda

That is all you should need in terms of getting software installed.

Main thing you are going to run into is RAM limitation on your laptop. There is no way around that. If your hardware does not support it then you have to rely on the cluster for those analyses.

[Dunois]

RAM, CPU (cores + speed), and storage space I'd say.

That said, OP, if you're looking for the easy way out, what you're looking for is Anaconda (<- clickable link). One shot installation of conda and all its data science friends (including IDEs and the likes). That'll get you python, numpy, etc., that you mentioned.