tomsing1

# This is an example of the setup function used to filter sensitive data from 
# recorded API responses generated by the vcr R package

library("vcr") # *Required* as vcr is set up on loading

# escapeRegex function from the Hmisc package
escape_regex <- function(x) {
  gsub("([.|()\\^{}+$*?]|\\[|\\])", "\\\\\\1", x)
}

tomsing1

#!/usr/bin/env bash
# This script retrieves BAM files from AWS and runs the samtools and somalier tools.
# Requirements:
# - AWS CLI with credentials
# - samtools
# - docker
# - paths.txt file with path to BAM files on AWS S3, one per line

set -e
set -x

tomsing1

#!/usr/bin/env bash

# This script creates QC reports for all FASTQ files found in the 'fastq' directory,
# aligns the (paired-end) reads to the references index & attempts to stitch overlapping
# read pairs into single sequences.

set -e
set -x
set -o pipefail
set -o nounset

tomsing1

# Examples are modified from the vtree vignette: https://cran.r-project.org/web/packages/vtree/vignettes/vtree.html

library(vtree)
library(dplyr)
library(gtsummary)

class(Titanic)  # array with counts
titanic <- crosstabToCases(Titanic)
head(titanic)  # data.frame with individual outcomes

tomsing1

This document shows how to mount an AWS S3 bucket on Mac OS X using goofyfs.

The first three steps illustrate how to use goofys

1. Install goofyfs via brew

brew cask install osxfuse
brew install goofys

tomsing1

---
title: "Introduction to the stringr package"
author: "Thomas Sandmann"
date: "7/10/2020"
output: html_document
editor_options: 
  chunk_output_type: console
---

```{r setup, include=FALSE}

tomsing1

# A closer look at the HTODemux() function

1. Centered log-ratio (CLR) normalization of the counts
2. Define K as the number of barcodes + 1
3. Run k-means / k-median clustering
4. Calculate mean (normalized) barcode expression within each cluster
5. For each barcode
  - identify the cluster with the lowest mean counts
  - fit a negative binomial distribution to the raw counts of each cluster
  - define the 99th percentile of the (background) distribution as threshold

tomsing1

library(flexmix)
library(countreg)

# poisson

## two clusters of 1000 measurements each
y1 <- rpois(n = 1000, lambda = 20)
y2 <- rpois(n = 1000, lambda = 1)

y <- c(y1, y2)

tomsing1

#' RLA plot
#' 
#' @param df data.frame
#' @param x Scalar character, column of `df` used for the x-axis
#' @param y Scalar character, column of `df` used for the y-axis
#' @param feature Scalar character, column of `df` identifying the feature identifier
#' @param facet_rows Character vector, column(s) of `df` used facet rows
#' @param facet_cols Character vector, column(s) of `df` used facet columns
#' @param group_cols Character vector, column(s) of `df` that define the
#' grouping variable for median centering.

tomsing1

#' Adjust voomed expression based on a limma fit
#' 
#' When a design that contains both covariates of interest and nuisance
#' covariates (e.g. batch terms) is fit with limma's `lmFit` function,
#' it is helpful to examine the effect of the adjustment, e.g. by 
#' performing MDS on the adjusted results. The `limma::removeBatchEffect`
#' could be used, but performs the full fitting procedure. Alternatively
#' The coefficients for the nuisance terms from the full model can be 
#' extracted from the fit and used to correct the input matrix (usually
#' log transformed abundances, e.g. from `limma::voom`).