ChIP-Seq Analysis – Using CLCGenomics Workbench Nov 16,2017 Ansuman Chattopadhyay, PhD Health sciences library system University of pittsburgh ansuman@pitt.edu
Transcription Factor ChIP-Seq Histone ChIP-Seq ATAC-Seq Topics Transcription Factor ChIP-Seq Histone ChIP-Seq ATAC-Seq
www.hsls.libguides.com/chipseq
Transcription Factor and Histone ChIP-Seq
ATAC-Seq Study
Galaxy : http://galaxy.crc.pitt.edu:8080/ Graphical User Interface based software Galaxy : http://galaxy.crc.pitt.edu:8080/ CLC Genomics Workbench
Software @ HSLS MolBio http://hsls.libguides.com/molbio/licensedtools/resources
NGS Software @ HSLS MolBio NGS Analysis Sanger Seq Analysis Human , Mouse and Rat NGS Analysis
CLCbio Genomics Workbench System Requirements Windows Vista, Windows 7, Windows 8, Windows 10, Windows Server 2008, or Windows Server 2012 Mac OS X 10.7 or later. Linux: Red Hat 5.0 or later. SUSE 10.2 or later. Fedora 6 or later. 8 GB RAM required 16 GB RAM recommended 1024 x 768 display required 1600 x 1200 display recommended Intel or AMD CPU required Minimum 10 GB free disc space in the tmp directory
CLC Plugins to Install CLC Workbench Client Plugin Histone ChIP-Seq Advanced Peak Shape Tools Plugin – Beta Download available at Top Right Corner
Integrating with the CLCbio Genomics Server @ CRC http://core.sam.pitt.edu/CLCBioServer
You need Secure Remote Access via Pulse to run CLCGx from off campus locations / Pitt Wireless
CLC files at the CRC HTC Cluster Reference Sequences Look for Folders organized by PI’s name
Create Folders at CRC-HTC
Create Folder in SaM-HTC Cluster 1 2
Create Workshop Folder@ FRANK 1 2 3
ChIP-Seq Workflow
Dataset https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE63716
GEO Dataset https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE63716
Download FASTQ Reads MyoD_Undiff_ChIP-Seq
Download FASTQ Reads MyoD_Undiff_ChIp-Seq
ENA : Download FASTQ Reads MyoD_Undiff_ChIp-Seq
Import : FASTQ Reads MyoD_Undiff_ChIp-Seq 1
Import : FASTQ Reads MyoD_Undiff_ChIp-Seq (single)
GEO Dataset – ATAC-Seq
STEP 1: Import Reads to CLC (Paired End) 2
STEP 1: Import Reads to CLC (Paired End) 3 4 5
FASTQ format http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2847217/
FASTQ Reads
FASTQC Project http://www.bioinformatics.babraham.ac.uk/projects/fastqc/
Step 2: Create a Seq QC Report 1 2
Trim Reads – Adapter Seq etc.
Create Adapter List
Create Adapter List
Create FAST QC Report
FASTQC Report
Read Mapping to Ref Genome http://www.ensembl.org/info/data/ftp/index.html
Read Mapping to Ref Genome
Read Mapping to Ref Genome
Read Mapping to Ref Genome
Read Mapping to Ref Genome
Read Mapping around GM20652 Result from MyOD1 ChIP-Seq
Peak Calling Strino etal.,BMC Bioinformatics, June 2016
Peak Calling Strino etal.,BMC Bioinformatics, June 2016 Landt etal.,Genome Research,2012
Peak Calling Strino etal.,BMC Bioinformatics, June 2016
Discovering Obvious Peaks The CLC shape-based peak caller finds peaks by building a Gaussian filter based on the mean and variance of the fragment length distribution, which are inferred from the cross-correlation profile Strino etal.,BMC Bioinformatics, June 2016
Peak Shape Score The Peak Shape Score is standardised and follows a standard normal distribution, so a p-value for each genomic position can be calculated as p-value=Φ(−Peak Shape Score of the peak centre), where Φ is the standard normal cumulative distribution function. Score = genomic coverage * filter; *: cross-correlation operator Score indicates how likely a genomic position is to be a center of a peak Strino etal.,BMC Bioinformatics, June 2016
Once the positive and negative regions have been identified, Peak Shape Filter Once the positive and negative regions have been identified, the CLC shape-based peak caller learns a filter that matches the average peak shape, which is called Peak Shape Filter. Strino etal.,BMC Bioinformatics, June 2016
Peak Shape Filter Strino etal.,BMC Bioinformatics, June 2016
Peak Detection peaks are called by first identifying the genomic positions whose p-value is higher than the specified threshold and which do not have any higher value in a window around them. The size of this window is determined by the filter as the longest distance between two positive values in the filter. These maxima define the center of the peak, while the peak boundaries are identified by expanding from the center both left and right until either the score becomes 0 or the peak touches a window boundary Strino etal.,BMC Bioinformatics, June 2016
Call Peaks using Peak Shape information
Call Peaks using Peak Shape information
Call Peaks using Peak Shape information
Peak Calls Result
Peak Calls Result
Annotate Peaks with near by genes
Annotate Peaks with near by genes
5Prime and 3Prime Gene Distance
ChIP-Seq Result
Compare Datasets
Compare Datasets
Compare Datasets
Compare Datasets
Commonly Used Open-Source Tool https://pypi.python.org/pypi/MACS2
Comparison of CLC Results with MACS2.0
Histone ChIP-Seq Li etal., Cell 2007.01.015
Histone ChIP-Seq
Histone Modifications Li etal., Cell 2007.01.015
Running Histone ChIP-Seq Classify Regions of variable length by Peak Shape
Running Histone ChIP-Seq
Running Histone ChIP-Seq
Running Histone ChIP-Seq
Histone ChIP-Seq Result
Histone ChIP-Seq Result Classified Gene Regions in the genome
H3K4Me3 – Diff : Result by Txnfactor ChIP-Seq tool
ATAC-Seq
ATAC-Seq Data Analysis
Comparison of DNAse-Seq Results
HSLS-MBIS and Genomics Analysis Core GAC Ansuman Chattopadhyay, PhD 412-648-1297 ansuman@pitt.edu Uma Chandran, PhD, MSIS 412-648-9326 Chandran@pitt.edu Sri Chaparala srichaparala@pitt.edu Carrie Iwema, PhD, MLS 412-383-6887 iwema@pitt.edu http://hscrf.pitt.edu/
Thanks To…. CLCBio Center for Research Computing Shawn Prince HSLS Sri Chaparala Carrie Iwema David Leung Michael Sweezer CLCBio Shawn Prince Center for Research Computing Mu Fangping