daskelly

Migrating from bitbucket to github

When migrating a repo from bitbucket to github, I like to take a simple and straightforward approach:

git remote set-url origin git@github.com:USERNAME/REPONAME.git

This example assumes that a remote branch named origin points to bitbucket, and you are resetting that URL to now point to github. Then you can git push to github.

daskelly

curl -O -L https://github.com/iontorrent/TS/archive/TorrentSuite_5.2.2-sp1.tar.gz

daskelly

#!/bin/bash

#PBS -N name
#PBS -l nodes=1:ppn=1,mem=10gb,walltime=6:00:00
#PBS -m n
#PBS -j oe
#PBS -r n

module load R/3.3.2 gcc/4.9.2
module load pandoc/1.13.1

daskelly

mini <- matrix(rnorm(4000*50), nrow=4000, ncol=50)
system.time(M <- cor(t(mini), t(mini), method='pearson'))

fastcorr <- function(x) {
	# fast correlation between rows of x
	# from http://stackoverflow.com/questions/18964837/fast-correlation-in-r-using-c-and-parallelization
	if (!is.matrix(x)) x <- as.matrix(x)
	x <- x - rowMeans(x)
	x <- x/sqrt(rowSums(x^2))
	tcrossprod(x)

daskelly

library(tidyverse)
library(Seurat)
# Find specific markers for each cell type

celltypes <-   # list of cell types (unique ident labels of Seurat object)
obj <-         # Seurat object
specific_markers <- NULL

# First we do all pairwise comparisons and retain any markers that 
# are even somewhat higher in the cell type of interest

daskelly

#!/bin/sh

# works on a machine running CentOS 7.7
# no error checking and this requires manual intervention. This script just gives 
# an overview of the process.
#
# https://github.com/rust-lang/cargo/issues/713 was helpful...

mkdir -p apps/mosh
cd apps/mosh

daskelly

# Intalling leafcutter on a CentOS 6.5 machine
# This machine has glibc 2.12 and leafcutter needs glibc 2.14
module purge
module load gcc/4.9.2

# Install glibc 2.14
mkdir $HOME/apps/glibc_install
cd $HOME/apps/glibc_install 
wget http://ftp.gnu.org/gnu/glibc/glibc-2.14.tar.gz
tar zxvf glibc-2.14.tar.gz

daskelly

library(tidyverse)
library(biomaRt)

genes <- # FILL ME IN ... Should be a vector of gene names
markers <- # FILL ME IN ... Should be a data.frame with at least one column called mgi_symbol

ensembl <- useMart("ensembl", dataset="mmusculus_gene_ensembl")  # Mouse
# GO:0009986 is cell surface
go_cellsurf <- "GO:0009986"
gene.data <- getBM(attributes=c("mgi_symbol", "ensembl_gene_id", "go_id"),

daskelly

library(tidyverse)
library(Seurat)
library(doParallel)
# Find specific markers for each cell type

n.cores <- 10
registerDoParallel(cores=n.cores)

celltypes <-   # list of cell types (unique ident labels of Seurat object)
obj <-         # Seurat object

daskelly

read10x <- function(dir) {
    names <- c("barcodes.tsv", "features.tsv", "matrix.mtx")
    for (i in 1:length(names)) {
        R.utils::gunzip(paste0(dir, "/", names[i], ".gz"))
    }
    file.copy(paste0(dir, "/features.tsv"), paste0(dir, "/genes.tsv"))
    mat <- Seurat::Read10X(dir)
    file.remove(paste0(dir, "/genes.tsv"))
    for (i in 1:length(names)) {
        R.utils::gzip(paste0(dir, "/", names[i]))