Skip to content

Latest commit

 

History

History
136 lines (107 loc) · 6.01 KB

transcriptomics.md

File metadata and controls

136 lines (107 loc) · 6.01 KB
Raw

Bioinformatics training: transcriptomics

Protocol

We will follow the protocol described in Tophat2 bioinformatic protocol published in Nature protocol, 2012 for this course.

Trans_pipeline

For simulated dataset:

  • Tophat2 ---> cufflinks ---> cuffmerge ---> cuffdiff ---> cummeRbund
  • Tophat2 ---> cuffdiff ---> cummeRbund
  • Tophat2 ---> featureCounts ---> DESeq2/edgeR

For real life dataset

(Just because HISAT2 is new and much faster than Tophat2):

  • HISAT2 ---> StringTie ---> Ballgown (There is an option to follow something similiar to cuffmerge as well)
  • HISAT2 ---> cufflinks pipeline (with or without cuffmerge) ---> cummeRbund
  • HISAT2 ---> featureCounts ---> DESeq2/edgeR
  • HISAT2 worklow pipeline Nature protocol publication

Day 1

  • Introduction of transcriptomic analyses

  • Discusssion regarding reference-based and de novo approaches

  • How to use tophat v2 - based on the Nature protocol (2012) paper

  • Download genome reference and create an index using bowtie2

  • in /work/projects/nn9305k/home//transcriptome/ref

$ ln ../../../../bioinf_course/transcriptomics/ref/Dm.BDGP6.dna.toplevel.fa .
$ ls ../../../../bioinf_course/transcriptomics/ref/Dm.BDGP6.91.gtf .
$ bowtie2-build Dm.BDGP6.dna.toplevel.fa Dm_BDGP6_genome
  • USAGE
$ bowtie2-build <Reference_fasta_file_name> <bowtie2-build_ref_index_name_that_you_will_use_later>

To do

  • Align the given reads to the Drosophila genome using tophat v2
  • create a new folder called tophat in /work/projects/nn9305k/home//transcriptome/
  • create a slurm script with time=12:00:00, ntasks=16 and mem-per-cpu=12Gb
$ tophat -p 16 -G ../ref/Dm.BDGP6.91.gtf -o <tophat_output_folder_name> ../ref/Dm_BDGP6_genome <read1> <read2>

Day 2

  • Discuss results from tophat v2 alignemnt
  • Read about cufflinks and differential expression analysis pipeline

To do

  • Run cufflinks on the tophat output
  • Within tophat folder create a slurm script with time=12:00:00, ntasks=16 and mem-per-cpu=12Gb
$ cufflinks -p 16 -o <cufflinks_output_folder_name> <tophat_output_folder_name>/accepted_hits.bam

Day 3

  • Discuss about (long) using tophat -> cufflinks -> cuffmerge -> cuffdiff pipeline (To find novel transcripts and genes)
  • Discuss about (short) using tophat -> cuffdiff pipeline (To calculate differential expression for only known genes and transcripts)

To do

  • Run cuffmerge and cuffdiff on Day 2's cufflinks output
  • create a text file and call it assemblies.txt and it should contain the information below
$ cat assemblies.txt
<cufflinks_output_folder_name_for_Con1_Rep1>/transcripts.gtf
<cufflinks_output_folder_name_for_Con1_Rep2>/transcripts.gtf
<cufflinks_output_folder_name_for_Con1_Rep3>/transcripts.gtf
<cufflinks_output_folder_name_for_Con2_Rep1>/transcripts.gtf
<cufflinks_output_folder_name_for_Con2_Rep2>/transcripts.gtf
<cufflinks_output_folder_name_for_Con2_Rep3>/transcripts.gtf
  • Within tophat folder create a slurm script with time=12:00:00, ntasks=8 and mem-per-cpu=12Gb
$ cuffmerge -o <cuffmerge_output_folder_name> -g ../ref/Dm.BDGP6.91.gtf -s ../ref/Dm.BDGP6.dna.toplevel.fa -p 8 assemblies.txt
  • For the long pipeline, create a slurm script
cuffdiff -o cuffdiff_long_output -p 16 -L Con1_l,Con2_l <cuffmerge_output_folder_name>/merged.gtf Con1_Rep1_tophat/accepted_hits.bam,Con1_Rep2_tophat/accepted_hits.bam,Con1_Rep3_tophat/accepted_hits.bam Con2_Rep1_tophat/accepted_hits.bam,Con2_Rep2_tophat/accepted_hits.bam,Con2_Rep3_tophat/accepted_hits.bam
  • For the short pipeline, create a slurm script
cuffdiff -o cuffdiff_short_output -p 16 -L Con1_s,Con2_s ../ref/Dm.BDGP6.91.gtf Con1_Rep1_tophat/accepted_hits.bam,Con1_Rep2_tophat/accepted_hits.bam,Con1_Rep3_tophat/accepted_hits.bam Con2_Rep1_tophat/accepted_hits.bam,Con2_Rep2_tophat/accepted_hits.bam,Con2_Rep3_tophat/accepted_hits.bam
  • Check the above two scripts and identify the difference

Day 4

  • Load cuffdiff output from short and long pipeline in R using cummeRbund

To do

  • Link to cummeRbund manual: Manual

Day 5

  • Introduction to counting reads

To do

  • Link to featureCounts manual: Manual

Day 6-7

  • DESeq2 and cummeRbund

To do

Day 8

Also, one can get the ortholog information from ENSEMBL

Day 9