Package release: lncRScan-TFBS (version 0.0.1, Jul 2016)
Author: Sun Lei
Email: leisuncumt(at)yahoo.com
Description: lncRScan-TFBS is a Perl package for analyzing transcription factor binding sites (TFBSs) of
long non-coding RNAs(lncRNAs), compared with protein-coding genes.
Licence: lncRScan-TFBS is distributed under the licence of GNU GENERAL PUBLIC (GPL) Version 3.
For other programs, libraries or modules called in lncRScan-TFBS, please refer to their licences respectively.
It is free to use lncRScan-TFBS for academic usage. If you use it for commercial activities,
please contact the author by email.
Copyright (C) 2016 Yangzhou University
---------------------------------------------------------------------------------------------------------
Contents
--------
1. Introduction
2. Dependencies
3. Package components
3. Installation
5. Usage examples
1. Introduction
---------------
lncRScan-TFBS is a Perl package for analyzing transcription factor binding sites (TFBSs) of
long non-coding RNAs(lncRNAs), compared with protein-coding genes. Based on Linux OS and several
third-part programs and modules, the package can be used to further analyze the peaks from peak
calling programs working on ChIP-Seq data. lncRScan-TFBS can conduct basic statistical analysis
on the ChIP-Seq peaks, finding the maximum likelihood TSS-peak pairs for either lncRNA or
protein-coding genes. And lncRScan-TFBS can conduct motif finding or enrichment analysis by
calling programs such as MEME, Weeder and PscanChIP.
2. Dependencies
---------------
lncRScan-TFBS depends on several third-part programs and modules:
1) Perl 2.7 -- for interpreting Perl scripts.
2) R programming -- for statistical analysis. (http://www.r-project.org/)
3) ggplot2 -- for graphic plotting. (http://ggplot2.org/)
4) Statistics::R -- Perl interface for using R programming. (https://metacpan.org/pod/Statistics::R)
5) Bio::Das -- Perl module for connecting Distributed Sequence Annotation System (DAS) of University of
California Santa Cruz (UCSC) genome browser (http://www.genome.ucsc.edu/cgi-bin/das/dsn).
6) Linux shell commands -- used in the package.
7) Weeder 2.0 (http://159.149.160.51/modtools/downloads/weeder2.html) - for de novo motif finding.
8) memechip (MEME SUITE 4.11.2, http://meme-suite.org/) -- for de novo motif finding.
9) PscanChIP 1.2 -- scan sequences for finding motif enrichment using motif models from JASPAR or TRANSFAC.
(http://159.149.160.88/pscan_chip_dev/)
2. Package components
---------------------
--README
--GPL_LICENCE.txt
--ChangeLog (records of version changes)
--scripts (core Perl scripts of lncRScan-TFBS)
--example (example files for using lncRScan-TFBS)
--peak.sorted.bed (sorted annotation of ChIP-Seq peaks, genomce build: mm10)
--PCT.sorted.bed (sorted annotation of protein-coding transcripts in BED format, genomce build: mm10)
--LNCT.sorted.bed (sorted annotation of lncRNA transcripts in BED format, genomce build: mm10)
3. Installation
---------------
Currently lncRScan-TFBS can be installed on Linux OS by following steps:
(1) extract the compressed source package by running
$ tar zxvf lncRScan-TFBS.tar.gz
$ cd lncRScan-TFBS/scripts
$ chmod +x *.pl
and then get current directory by
$ pwd
(2) Add the path of lncRScan-TFBS scripts (got by 'pwd' above) to the environment variable $PATH
by modifying ~/.bashrc in your home directory:
First, in the end of the file add
export PATH=$PATH:<the directory you have got by 'pwd' in previous step>
then save ~/.bashrc and restart the shell or running "$ source ~/.bashrc".
After that you can use the scripts of lncRScan-TFBS on the shell.
(3) All dependent third-part packages, libraries and modules should be installed and
the directory of their executable binaries should be set in the environment variable $PATH.
5. Usage Example
-----------------
The package lncRScan-TFBS can be used to analyse the TFBSs step by step as follows (Skip 1 and 2 for the example):
(1) Preparing three input files
peak.bed -- A BED file of ChIP-Seq peak loci determinded by peak calling programs such as MACS with fields:
"chrom chromStart chromEnd name score strand"
LNCT.bed -- A BED file of lncRNA genes/transcripts
PCT.bed -- A BED file of protein-coding genes/transcripts
Note: LNCT denotes long non-coding transcript and PCT is for protein-coding transcript.
The comprehensive usage of each lncRScan-TFBS script can be seen by "$ command.pl -h|--help".
(2) Sorting input files
$ sort -k1,1 -k2n peak.bed > peak.sorted.bed
$ sort LNCT.bed > LNCT.sorted.bed
$ sort PCT.bed > PCT.sorted.bed
Note: If you only have GTF files of gene annotation, please try to convert GTF to BED format using
gtf2bed.py of lncRScan-SVM (http://sourceforge.net/projects/lncrscansvm/?source?directory)
developed by our group.
(3) Statistics of input files using
$ pre_stats.pl -c peak.sorted.bed -l LNCT.sorted.bed -p PCT.sorted.bed
The results include four files in a folder named pre_stats_out in current directory.
The first one is counts.output containing the counts of peak loci, LNCTs and PCTs respectively.
The second and third ones are peak_length.summary and peak_value.summary, which show several
basics statistics such as mean and median of the peak lengths and values respectively. And the last
peak_plots.pdf shows distributions of peak lengths and values with density plots, and a scatter plot
is for viewing the relationship between peak length and value.
(4) TSS-peak pairs and statistics
For each transcript, find the peak locus nearby which has the shortest distance
within 3000 bp upstream its TSS.
$ tss_peak_distance.pl -g LNCT.bed -c peak.bed -m min_dis -u 5000 -o LNCT_peak
$ tss_peak_distance.pl -g PCT.bed -c peak.bed -m min_dis -u 5000 -o PCT_peak
Or find the peak locus nearby which has the highest peak value within 3000 bp upstream its TSS.
$ tss_peak_distance.pl -g LNCT.bed -c peak.bed -m max_value -u 5000 -o LNCT_peak
$ tss_peak_distance.pl -g PCT.bed -c peak.bed -m max_value -u 5000 -o PCT_peak
The programs above will search available peaks within 5000 bp upstream TSS of each transcript
and finally output four files, including LNCT_peak.table, PCT_peak.table, LNCT_peak.bed and PCT_peak.bed.
LNCT_peak.table and PCT_peak.table contains information of TSS-peak pairs while LNCT_peak.bed
and PCT_peak.bed are the non-redundant peak annotion of the TSS-peak pairs.
Then, we can conduct basic statistical computation and visualization
on the resulting TSS-peak pairs.
$ tss_peak_stats.pl -l|--lnc LNCT_peak.table -p|--pc PCT_peak.table
This program will calculate distributions of TSS-peak distances and peak values
of LNCTs and PCTs respectively and some statistics, and output eight files in a
folder named tss_peak_stats_out. The output counts.output shows the count of TSS-peak pairs.
The output files test.distance.output and test.value.output containing results of
Wilcoxon tests on the TSS-peak distances and values respectively, and some statistics
are shown in four summary files. The output tss_peak_plots.pdf shows distributions
of the distances and peak values between PCTs and LNCTs, and relationships between
them in scatter plot.
(5) De novo motif finding
In this step, we first fetch genomic sequences corresponding to the peak loci chose
by previous step from UCSC DAS server using fetch_seq_bed.pl.
$ fetch_seq_bed.pl -i PCT_peak.bed -g mm10 -o PCT_peak.fa
$ fetch_seq_bed.pl -i LNCT_peak.bed -g mm10 -o LNCT_peak.fa
Then, search for potential TFBS patterns using pattern_search.pl, which can call meme-chip or weeder.
You can use memechip:
$ motif_finding.pl -i PCT_peak.fa -s MM -t memechip -o memechip_out_PCT
$ motif_finding.pl -i PCT_peak.fa -s MM -t memechip -o memechip_out_LNCT
or use weeder2:
$ motif_finding.pl -i PCT_peak.fa -s MM -t weeder2 -o weeder_out_PCT
$ motif_finding.pl -i PCT_peak.fa -s MM -t weeder2 -o weeder_out_LNCT
All output files created by meme-chip or weeder2 will be put in the output directories.
You can setting avaliable parameters for using memechip or weeder2 by modifying the script motif_finding.pl.
(6) Motif enrichment analysis
In this step, the peak loci of PCTs of LNCTs are scanned using TFBS models of JASPAR or TRANSFAC by PscanChIP.
$ motif_enrich.pl -p PCT_peak.bed -l LNCT_peak.bed -g dir_to_genome -M dir_to_PscanChIP/jaspar_2016.wil -bg dir_to_background -o out_folder
where dir_to_genome is the directory containing raw genome sequences defined by PscanChIP, dir_to_PscanChIP/jaspar_2016.wil
is always in the folder of PscanChIP package, and dir_to_background is always in a subdirectory named 'BG' of the home directory of PscanChIP package downloaded. Please refer to the documentation of PscanChIP for setting parameters.
The intermediate and final output files will be all put in out_folder. Specifically, TF_both_sig, TF_PCT_sig and TF_LNCT_sig will list TFs showing significant enrichment in both samples, PCT samples and LNCT samples respectively.
