TopHat 2014.10.01 Mi-kyoung Seo. Today’s paper..TopHat Cole Trapnell at the University of Washington's Department of Genome Sciences Steven Salzberg Center.

TopHat

2014.10.01Mi-kyoung Seo

Today’s paper..TopHat

Cole Trapnellat the University of Washington's Department of Genome Sciences

Steven Salzberg Center for Computational Biology at Johns Hopkins University

Pipeline for RNA-Seq

TopHat

FastQCData quality controlData quality control

MappingMapping

Differential expression analysis

Differential expression analysis

Cufflinks(Cuffdiff)

FPKM for genes (transcripts)

Cleaned reads

Mapped reads

R DEG

Transcripts assemblyTranscripts assembly

Final transcripts assembly

Final transcripts assembly

8 RNA-seq2x100bp readsSequencingSequencing

Tuxedo protocol2012, Nature Method

Software information

• Purpose– A spliced read mapper for RNA-Seq

• TopHat has a number of parameters and options, and their default values are tuned for processing mammalian RNA-Seq reads

• Software URL– http://ccb.jhu.edu/software/tophat/index.shtml

• Category– Aligner

• License– Open source and freely available under the Artistic license

Overview of RNA-Seq

5Garber, M., et al. (2011), Nature Methods, 8(6), 469–477.Microarray or cDNA/EST sequencing RNA-seq

QPALMA

• De Bona et al., 2008• Optimal spliced alignments of short sequence reads• Q-PALMA is based on a machine learning approach, in which

data from previously known splice junctions are used to train the software.

• Initial mapping phase uses Vmatch (Abouelhoda et al., 2004)• Vmatch is a flexible, fast aligner, but because it is not

designed to map short reads on machines with small main memories, it is substantially slower than other specialized short-read mappers.– Vmatch: 180,000 reads per CPU hour

– TopHat: 2.2 million reads per CPU hour

The Tophat pipeline

1. Find Exons- Mapping using BowtieMapped & IUM reads- Assemble exons using MAQPutative exons (islands)- Extend exons by 45bp- Gap

2. Find Splice Junctions

Mapping: Bowtie

TopHat uses Bowtie in order to initially map all reads to the reference genome while collecting all the unmapped reads for further analysis

Bowtie is reporting reads with no more than a few mismatches (default: 2) within s bp from the 5' end and the 3' end may have errors based on Phred-Quality-Weighted Hamming Distance (s default: 28)

TopHat allows reads from bowtie that map up to 10 locations (multireads)

read

s bp

5' 3'

Assembly: MAQ

Exon 2 Exon 3Exon 1

read

read

Exon 2Exon 1Transcript 1

Gene(DNA or reference)

read read

GU AG

read

read

readread

readread

read

readread

readread

island

consensus

read

read

Putative exons

IUM reads

Identification of read spanning junction

Exon 2 Exon 3Exon 1

read

GT AG AG AG AG

70 < potential Intron length < 20,000bp

IUM reads

Identification of read spanning junction

ERAGNE (annotation based pipeline)(Mortazavi et al., 2008)

Each island spanning coordinates i to jdm: depth of coverage at coordinate mn: the length of the reference genome

B3gat1 gene

Seed and extend strategy

To find reads that span junction

2K-mer seed

Running Operations

• Input– For Paired-end read, Read1.fastq, Read2.fastq– genome (Genome index by bowtie-build)– Option: genes.gtf (Gene annotation)

• Command– Usage: tophat [options]* <genome_index_base>

<reads1_1[,...,readsN_1]> [reads1_2,...readsN_2]– tophat –p 4 –o output/Normal –G genes.gtf BowtieIndex/genome

read1.fastq read2.fastq --library-type fr-unstranded• TopHat Output

– accepted_hits.bam– unmapped.bam– junctions.bed– insertion.bed– deletion.bed

Usage: tophat [options]* <genome_index_base> <reads1_1[,...,readsN_1]> [reads1_2,...readsN_2]$ tophat –p 4 –o output/Normal –r 100 –G genes.gtf BowtieIndex/genome read1.fastq read2.fastq

Input Parameters

• $ tophat [optioins] <bowtie_index_base> <read1_1> <read1_2>

• -o: output dir• -p: use this many threads to align reads [default: 1]• -r: this is the expected (mean) inner distance between mate

pairs [default: 50bp]– For, example, for paired end runs with fragments selected

at 300bp, where each end is 50bp, you should set -r to be 200. The default is 50bp

• -G: geneset

$ tophat –p 4 –o output/Normal –r 100 –G genes.gtf BowtieIndex/genome read1.fastq read2.fastq

Tophat output by IGV

Discussion

• Tophat2 default option 확인– r: 실험에 따라 다름– Bowtie2 for mapping