millerh1

---
title: "DGE + Pathway Analysis for Minimalists"
author: "Henry Miller"
date: "07/22/2021"
output:
  html_document:
    toc: true
    toc_float:
      collapsed: false
      smooth_scroll: false 

millerh1

env:
  RENV_PATHS_ROOT: ~/.local/share/renv
  
on:
  push:
    branches:
      - main
  pull_request:
    branches:
      - main

millerh1

UCSC_orgID	description	nibPath	organism	defaultPos	active	orderKey	genome	scientificName	htmlPath	hgNearOk	hgPbOk	sourceName	taxId	genes_available	year	eff_genome_size_36bp	eff_genome_size_50bp	eff_genome_size_75bp	eff_genome_size_100bp	eff_genome_size_125bp	eff_genome_size_150bp	eff_genome_size_200bp	eff_genome_size_250bp	eff_genome_size_300bp	genome_length	homer_anno_available
ailMel1	Dec. 2009 (BGI-Shenzhen 1.0/ailMel1)	/gbdb/ailMel1	Panda	GL192818.1:558576-566855	1	16070	Panda	Ailuropoda melanoleuca	/gbdb/ailMel1/html/description.html	0	0	BGI-Shenzhen AilMel 1.0 Dec. 2009	9646	TRUE	2009	2131884815	2195257147	2167233843	2206501040	2213941269	2196290908	2233702932	2212959190	2168863899	2299509015	FALSE
allMis1	Aug. 2012 (allMis0.2/allMis1)	/gbdb/allMis1	American alligator	JH731472:504271-884586	1	1425	American alligator	Alligator mississippiensis	/gbdb/allMis1/html/description.html	0	0	International Crocodilian Genomes Working Group	8496	TRUE	2012	2050346072	2080458921	2117005123	2101224009	2132575305	21480

millerh1

get_public_run_info <- function(accessions) {
  
  suppressWarnings(suppressPackageStartupMessages(require(XML)))
  
  httr::set_config(httr::config(http_version = 2))
  # set the HTTP version to 1.1 (none, 1.0, 1.1, 2)
  #### Bug testing ##
  #accessions <- c("SRX2481503", "SRX2481504", "GSE134101", "SRP150774", "GSE127329", "SRS1466492")
  #accessions <- c("SRX2918366", "SRX2918367", "GSM3936516", "SRX5129664")

millerh1

# Script for wrangling test files to correct naming conventions
library(tidyverse)
library(parallel)

S3_BAM_URI <- "s3://rseq-testing/bam-files/"
bamsAvail <- system(paste0("aws s3 ls ", S3_BAM_URI), intern = TRUE)

oldnew <- tibble(
  oldfls = gsub(bamsAvail, pattern = ".+ ([ES]{1}RX[0-9]+_.+\\.[hgmm]{2}[0-9]+\\.bam)", replacement = "\\1"),
  newfls = gsub(bamsAvail, pattern = ".+ ([ES]{1}RX[0-9]+)_.+\\.([hgmm]{2}[0-9]+\\.bam)", replacement = "\\1_\\2")

millerh1

#!/bin/bash

bamToBed -i $BAM_ORIG -bed12 -cigar | shuffleBed -i - -g $CHROM_SIZES | bedToBam -bed12 -g $CHROM_SIZES | samtools sort > $BAM_SHUFFLED

millerh1

encExp2Peaks <- function(acc, getIDR=FALSE) {
  resp <- httr::GET(url = paste0("https://www.encodeproject.org/search/?type=File&dataset=/experiments/",
                                 acc,"/&file_type=bed+narrowPeak&format=json&frame=object&limit=all"), httr::accept_json()) %>%
    httr::content(as="parsed")
  res <- lapply(seq(resp$`@graph`), function(i) {
    entry <- resp$`@graph`[[i]]
    tibble(
      accfile = entry$accession,
      sup = ! is.null(entry$superseded_by %>% unlist()),
      type = entry$file_type,

millerh1

# Create links in datatabels for gene info modal
createGeneInfoLink <- function(val) {
  sprintf(paste0('<a href="https://www.genecards.org/cgi-bin/carddisp.pl?gene=', val ,'" target="_blank" class="tooltip-test" onClick="gene_click(this.id)" title="', val, '">', val, '</a>'))
  
}
# Create links in datatables for GSEA info from MSIGDB
createGSEAInfoLink <- function(val, valTitle) {
  sprintf(paste0('<a href="http://software.broadinstitute.org/gsea/msigdb/cards/', val,'" id="', val ,'" target="_blank" class="tooltip-test" onClick="gsea_click(this.id)" title="', val, '">', valTitle, '</a>'))
}
# Create links in datatables for GEO accessions

millerh1

library(magrittr)

# Venn diagram plot in tidy pipe
plt <- list(
    "A" = 1:100,
    "B" = 96:140
) %>%
    VennDiagram::venn.diagram(filename = NULL) %>%
    grid::grid.draw(.) %>%
    ggplotify::grid2grob() %>% 

millerh1

library(msigdbr)
library(dplyr)
msigdbr(category = "C8") %>% 
  dplyr::select(gs_name, gene_symbol) %>% 
  distinct(gs_name, gene_symbol) %>% 
  mutate(val = 1) %>% 
  pivot_wider(
    id_cols = gs_name, names_from = gene_symbol, 
    values_from = val, 
    values_fill = 0