Tuxedo Suite For Splice Variant Analysis and Identifying Novel Transcripts II

All these steps have already been run. We'll be spending time looking at the commands and output. Let's get set up.

Get the data

Step 1: Run tophat2

We've already gone over how this is done and looked over the results. Let's move on to step 2.

Step 2: Run cufflinks

General syntax for cufflinks command

cufflinks [options] <hits.bam>

Some of the important options:

Look at $BI/ngs_course/tophat_cufflinks/run_commands/cufflinks.commands to see how it was run.

How do I look into the file?

cat $BI/ngs_course/tophat_cufflinks/run_commands/cufflinks.commands

Take a minute to look at the output files produced by one cufflinks run.

The important file is transcripts.gtf, which contains Tophat's assembled junctions for C1_R1.

Cufflinks output files

cd $BI/ngs_course/tophat_cufflinks/C1_R1_clout
ls -l


drwxrwxr-x 2 nsabell G-801021    32768 May 22 15:10 cuffcmp
-rwxr-xr-x 1 daras G-803889  14M Aug 16 12:49 transcripts.gtf
-rwxr-xr-x 1 daras G-803889 597K Aug 16 12:49 genes.fpkm_tracking
-rwxr-xr-x 1 daras G-803889 960K Aug 16 12:49 isoforms.fpkm_tracking
-rwxr-xr-x 1 daras G-803889    0 Aug 16 12:33 skipped.gtf

Step 3: Merging assemblies using cuffmerge

We first create a file listing the paths of all per-sample transcripts.gtf files so far, then pass that to cuffmerge:

How did we do that?

cd $BI/ngs_course/tophat_cufflinks
find . -name transcripts.gtf > assembly_list.txt
cuffmerge <assembly_list.txt>

assembly_list.txt contents

cat $BI/ngs_course/tophat_cufflinks/assembly_list.txt

Take a minute to look at the output files produced by cuffmerge. The most important file is merged.gif, which contains the consensus transcriptome annotations cuffmerge has calculated.

cuffmerge output

cd $BI/ngs_course/tophat_cufflinks/merged_asm
ls -l

-rwxrwxr-x  1 daras G-803889  1571816 Aug 16  2012 genes.fpkm_tracking
-rwxrwxr-x  1 daras G-803889  2281319 Aug 16  2012 isoforms.fpkm_tracking
drwxrwxr-x  2 daras G-803889    32768 Aug 16  2012 logs
-r-xrwxr-x  1 daras G-803889 32090408 Aug 16  2012 merged.gtf
-rwxrwxr-x  1 daras G-803889        0 Aug 16  2012 skipped.gtf
drwxrwxr-x  2 daras G-803889    32768 Aug 16  2012 tmp
-rwxrwxr-x  1 daras G-803889 34844830 Aug 16  2012 transcripts.gtf

Step 4: Finding differentially expressed genes and isoforms using cuffdiff

cuffdiff [options] <merged.gtf> <sample1_rep1.bam,sample1_rep2.bam> <sample2_rep1.bam,sample2_rep2.bam>

Exercise: What does cuffdiff -b do?

Solution

-b is for enabling fragment bias correction.