cufflinks result interpretation

Question

[priyadhawan11234]

i have 2 rna sample and when did cufflinks with aligned bam files , i have got result for sample1 in transcript.gtf file -

2       Cufflinks       transcript      29643815        29738975        1       +       .       gene_id "ENSMUSG00000039844"; transcript_id "ENSMUST00000238899"; FPKM "0.0325281086"; frac "0.000455"; conf_lo "0.012566"; conf_hi "0.052490"; cov "1.448822";
2       Cufflinks       exon    29643815        29643914        1       +       .       gene_id "ENSMUSG00000039844"; transcript_id "ENSMUST00000238899"; exon_number "1"; FPKM "0.0325281086"; frac "0.000455"; conf_lo "0.012566"; conf_hi "0.052490"; cov "1.448822";
2       Cufflinks       exon    29679133

for sample2 -

2       Cufflinks       transcript      29643815        29738975        163     +       .       gene_id "ENSMUSG00000039844"; transcript_id "ENSMUST00000238899"; FPKM "1.0183970793"; frac "0.071810"; conf_lo "0.851414"; conf_hi "1.185380"; cov "14.202789";
2       Cufflinks       exon    29643815        29643914        163     +       .       gene_id "ENSMUSG00000039844"; transcript_id "ENSMUST00000238899"; exon_number "1"; FPKM "1.0183970793"; frac "0.071810"; conf_lo "0.851414"; conf_hi "1.185380"; cov "14.202789";

both sample have same exon numbers but rpkm value is diffrenent . what interpretation can we get from this result. what is meaning of this result?

Answer 1

[Kevin Blighe]

You mean FPKM, not RPKM. The interpretation is that the expression is higher in sample 2 when compared to sample 1. That is all.

Cufflinks is out dated, by the way. You should be using HISAT2 / StringTie if your aim is to perform transciptome assembly. Otherwise, a pseudoaligner like Salmon or Kallisto should suffice.

Also, FPKM and RPKM units should not be used if your end-goal is to perform differential expression analysis.

Kevin

[ATpoint]

Grammar nazi here: Salmon in its current version now performs selective alignment :)

[priyadhawan11234]

what-else interpretation can we do from above output ? for example can we know transcript length and exon numbers in both samples from this output?

[Kevin Blighe]

That information is right there in your output. Please take another look.

[Kristoffer Vitting-Seerup]

Tophat(2) is outdated - Cufflinks is not...?

[Kevin Blighe]

I had thought that, generally, HISAT2 and StringTie were the upgrades to TopHat2 and Cufflinks, respectively... or am I wrong?

Answer 2

[Kristoffer Vitting-Seerup]

Please note that if you want to compare the expression of the two samples you need to first run Cuffmerge and then requantify the samples on the joint transcriptome. You can read more at cufflinks website or here.