Cufflinks workflows

[sheenams]

Hi,

I’m running a few workflows through Cufflinks/Cuffcompare/Cuffdiff in Galaxy (only because I can’t get it to install into my 32bit Ubuntu).
I have 3 types of analyses, but only one seems to be working. Can you help me understand what I am doing wrong?

I have 2 time points: CON (control) and REG (regenerated), 3 age cohorts (J-juvenille, A-adult, G-Geriatric), 3 biological replicates per time point/age (J-CON(1-3), J-REG(1-3), A-CON(1-3),etc).

I’m using the Ensembl gene track from UCSC for Zebrafish. The GTF looks like this:

chr1 danRer7_ensGene exon 25135496 25135824 0.000000 - . gene_id "ENSDART00000141301"; transcript_id "ENSDART00000141301";
chr1 danRer7_ensGene exon 25177126 25177291 0.000000 - . gene_id "ENSDART00000141301"; transcript_id "ENSDART00000141301";

Workflow 1: I’m trying to test each replicate against itself for differentially expressed genes (CON1 vs REG1).

1. Run Cufflinks on each sample, with the Ensembl GTF as the reference annotation.
2. Use the GTFs from cufflinks and Ensembl GTF in Cuffcompare
3. Use combined transcripts from cuffcompare, with TopHat output of CON and REG.

My output looks like this:

test_id gene_id gene locus sample_1 sample_2 status value_1 value_2 ln(fold_change) test_stat p_value q_value significant
XLOC_000001 XLOC_000001 ENSDART00000127092 Zv9_NA10:29174-34103 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000002 XLOC_000002 ENSDART00000130683 Zv9_NA101:39833-39948 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000003 XLOC_000003 ENSDART00000130536 Zv9_NA105:12113-12230 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000004 XLOC_000004 ENSDART00000123528 Zv9_NA109:103-187 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000005 XLOC_000005 ENSDART00000128305 Zv9_NA109:280-364 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000006 XLOC_000006 ENSDART00000123913 Zv9_NA109:457-541 CON REG NOTEST 0 0 0 0 1 1 no

That appears to be working right.

Workflow 2: I’m trying to test each age cohort CON vs REG against itself for differentially expressed genes (CON(1-3) vs REG(1-3)).
1. run Cufflinks on each sample, with the Ensembl GTF as the reference annotation.
2. Use the GTFs (total of 6) from cufflinks and Ensembl GTF in Cuffcompare
3. Use combined transcripts from cuffcompare, with TopHat output of CON and REG in Cuffdiff, with 2 groups of replicates (CON and REG)

test_id gene_id gene locus sample_1 sample_2 status value_1 value_2 ln(fold_change) test_stat p_value q_value significant
XLOC_000001 XLOC_000001 - Zv9_NA10:29174-34103 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000002 XLOC_000002 - Zv9_NA101:39833-39948 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000003 XLOC_000003 - Zv9_NA105:12113-12230 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000004 XLOC_000004 - Zv9_NA109:103-187 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000005 XLOC_000005 - Zv9_NA109:280-364 CON REG NOTEST 0 0 0 0 1 1 no
XLOC_000006 XLOC_000006 - Zv9_NA109:457-541 CON REG NOTEST 0 0 0 0 1 1 no

Notice how there is nothing in ‘gene’? There is nothing in that column other than ‘-‘ . How do I get it to output the Ensembl ID like in the first analysis?

Workflow 3: I’m trying to test each age cohort CON against the other age CON for differentially expressed genes (A-CON(1-3) vs G-CON(1-3) vs J-CON(1-3)).
• I ran this two different ways. One with 2 replicates ( A vs G, G vs J, A vs J) and one with all three replicates (A vs G vs J)
1. run Cufflinks on each sample, with the Ensembl GTF as the reference annotation.
2. Use the GTFs (total of 9 ) from cufflinks and Ensembl GTF in Cuffcompare
3. Use combined transcripts from cuffcompare, with TopHat output of CON Cuffdiff, with 2 groups of replicates (A-CON(1-3) vs G-CON(1-3) vs J-CON(1-3)) and with 3 groups : A vs G vs J

The output from 2 replicate groups

test_id gene_id gene locus sample_1 sample_2 status value_1 value_2 ln(fold_change) test_stat p_value q_value significant
XLOC_000001 XLOC_000001 - Zv9_NA10:29174-34103 A G NOTEST 0 0 0 0 1 1 no
XLOC_000002 XLOC_000002 - Zv9_NA101:39833-39948 A G NOTEST 0 0 0 0 1 1 no
XLOC_000003 XLOC_000003 - Zv9_NA105:12113-12230 A G NOTEST 0 0 0 0 1 1 no
XLOC_000004 XLOC_000004 - Zv9_NA109:103-187 A G NOTEST 0 0 0 0 1 1 no
XLOC_000005 XLOC_000005 - Zv9_NA109:280-364 A G NOTEST 0 0 0 0 1 1 no
XLOC_000006 XLOC_000006 - Zv9_NA109:457-541 A G NOTEST 0 0 0 0 1 1 no


The output from 3 replicate groups:

test_id gene_id gene locus sample_1 sample_2 status value_1 value_2 ln(fold_change) test_stat p_value q_value significant
XLOC_000001 XLOC_000001 - Zv9_NA10:29174-34103 G J NOTEST 0 0 0 0 1 1 no
XLOC_000002 XLOC_000002 - Zv9_NA101:39833-39948 G J NOTEST 0 0 0 0 1 1 no
XLOC_000003 XLOC_000003 - Zv9_NA105:12113-12230 G J NOTEST 0 0 0 0 1 1 no
XLOC_000004 XLOC_000004 - Zv9_NA109:103-187 G J NOTEST 0 0 0 0 1 1 no
XLOC_000005 XLOC_000005 - Zv9_NA109:280-364 G J NOTEST 0 0 0 0 1 1 no
XLOC_000006 XLOC_000006 - Zv9_NA109:457-541 G J NOTEST 0 0 0 0 1 1 no


Only the first workflow gives the Ensembl gene in the output. Can anyone please help me understand why? Am I doing the analysis wrong or do I need to merge this output file with another file to get the Ensembl gene names? Is there a better way of doing this analysis?

Thanks!