Software installation on Debian based system

Some suggested apt-get commands

Most software should be installed via conda especially for bioinformatics (see later). Here are a few system-wide software better installed via apt-get. They work for Debian 9 or distributions based on Debian 9 (have not tested them on Ubuntu):

sudo apt-get install git curl unzip gzip bzip2 ca-certificates build-essential gfortran libgfortran-6-dev libgomp1 libgsl-dev libatlas3-base liblapack-dev libcurl4-openssl-dev libssl-dev
sudo apt-get install openssh-server
sudo apt-get install vim vim-gtk3
sudo apt-get install trash-cli
sudo apt-get install texlive-full python-pygments fonts-wqy-microhei ttf-wqy-microhei fonts-wqy-zenhei ttf-wqy-zenhei pandoc graphviz libmagickwand-dev nodejs
sudo apt-get install smartmontools lm-sensors htop
sudo apt-get install dos2unix tree sshpass
sudo apt-get clean

Install latest R

I would not recommend it, at least at this point, to source compile R or to install R via conda (unless you have experiences with these in the past). Here we use the Linux distribution’s package repository to install.

Debian 9 upgrade

For R 3.5 on Debian 9, for example, create a file

sudo gvim /etc/apt/sources.list.d/cran.list

and add

deb stretch-cran35/

and save and exit. This specifies the path to the repository from which the Debian system gets the latest R version (R 3.5) for its distribution (Debian 9, aka, Debian Stretch). Then add GPG key to use the repository:

sudo apt-key adv --keyserver --recv-key 'E19F5F87128899B192B1A2C2AD5F960A256A04AF'

Ubuntu 14.04 upgrade

Add the line

deb trusty-cran35/

to /etc/apt/sources.list.d/cran.list, then

sudo apt-key adv --keyserver --recv-keys E084DAB9 

Finally, install it:

sudo apt-get update & sudo apt-get install r-base r-base-dev libatlas3-base

Some suggested conda commands for bioinformatics tools

conda install -c bioconda cyvcf2 bedtools plink vcftools bcftools bedops tabix htslib

But other than bioinformatics toos, I generally prefer to use pip to install Python packages; I prefer not using conda for either Python or R packages.