agotron_detector

Introduction

The agotron_detector toolkit identifies and quantifies agotrons in Ago CLIPseq datasets.

Prerequisites

bowtie2 (tested with v2.2.8)

samtools (tested with v1.3)

python (tested with v2.7.11) including modules: pysam, numpy, and mySQL

R (tested with v3.2.5) including packages: ggplot2, ggregex, dplyr, tidyr, and optparse

Scripts / Usage

The agotron_detector repository basically contains three scripts that should be run sequencially:

(1) UCSC_intron_retriever.py, a python script to extract short introns from the UCSC mySQL server

Usage:
  UCSC_intron_retriever.py [ARGUMENTS] > [OUTPUT]
  
Options:
  -db <string>      MySQL database (default='hg19')
  -table <string>   MySQL table (default='refGene')
  -max <int>        Maximum intron length (default=150)
  -min <int>        Minimum intron length (default=50)

Alternatively, use UCSC_mirna_retriever.py or tophat_intron_retriever.py to retrieve RefSeq annotated miRNA coordinates or intron coordinates from tophat-produced junctions.bed.

(2) analyzer.py, a python script to intersect mapped reads with coordinates of interest and output agotron-relevant features

Usage:
  analyzer.py [ARGUMENTS] < [INPUT] > [OUTPUT]
  
Required arguments:	
  -g <file>         Path to reference genome fastafile, must be indexed with samtools faidx
Optional arguments:	
  -f <files…>       Input bam-files (default=*.bam)
  -c <file>         Filename for coverage output (if empty, no coverage file is produced) 
                    (default='coverage.txt')
  -tr <int>         RPMM (reads per mapped million) expression threshold for output
                    (default = 5)
  -ts <int/'all'>   How many samples to meet RPMM expression threshold. For all samples, type ‘all’ 
                    (default = 2)
  -m <float>        Tolerance for reads mapping partly outside locus 
                    (default=0.1, e.g. 10% of the reads is allowed to map outside locus)
  -q <int>          Threshold for mapping quality 
                    (default=13)
  -a <int>          Add <int> flanking nucleotides (up and downstream) to the loci sequence output 
                    (default=10) 

(3) annotater.R, an R script that annotates agotron and outputs a few different plots

Usage: 
  annotater.R [ARGUMENTS] < [INPUT]	

optional arguments:	
  -c <file>         Input coverage file (the -c output from analyzer.py) 
                    (default='coverage.txt')
  -p <string>       Prefix used in output files 
                    (default='agotron')
agotron definition:	
  -m <int>          Threshold for median read length 
                    (default=30)
  -h <float>        Minimum fraction of reads with distance (-d) between 5’end of read and 5’end of locus 
                    (default=0.7)
  -d <int>          Maximum distance allowed from predominant 5’end of reads to 5’end of locus 
                    (default=1)

Example

Clone the repository and run the example script, all_commands.sh. This:

1. Downloads and prepares reference genome (hg19)
2. Downloads and trims Ago CLIPseq dataset (GSE78059)
3. Maps dataset to the reference genome using bowtie2
4. Intersects the mapped reads (bam-files) with annotated short introns to detect and annotate agotrons.

sh all_commands.sh

To annotate agotrons in mapped data (use -db and -g options to specify the reference genome used):

python UCSC_intron_retriever.py -db hg19 | python analyzer.py -g /path/to/hg19.fa -f /path/to/*.bam | Rscript annotater.R 

Citation

Hansen TB. Detecting agotrons in Ago CLIPseq data. MiMB, 2017, submitted

License

Copyright (C) 2017 ncRNALab. See the LICENSE file for license rights and limitations (MIT).