I merged two transcripts using cuffmerge that are de novo assembled by cufflinks. The resulting cuffmerge-transcript.gtf has very low FPKM values compared to cufflinks-transcripts.gtf files. More importantly the cuffmerge-transcripts.gtf file FPKM values are in between 0 and 1 range, where as cufflinks-transcripts 0 to 160,000 range. I used the following commands ...
Does any one know why cuffmerge is producing weird FPKM scores ?
cufflinks -o cufflinks_rep1 tophat_h1hesc_rep2/accepted_hits.bam cufflinks -o cufflinks_rep2 tophat_h1hesc_rep2/accepted_hits.bam assemblies.txt: cufflinks_rep1/transcripts.gtf cufflinks_rep2/transcripts.gtf cuffmerge -s hg19.fa assemblies.txt