Some quantifiers (e.g. Cufflinks) can produce vastly inflated FPKM values for very short transcripts, in which case it is recommended to simply ignore the values produced for such isoforms. Unfortunately piquant converts all FPKM values into TPMs for analysis, and in the case of very high FPKMs, this conversion leads to nearly all TPMs being close to zero - even though most calculated FPKMs are still reasonable.
It will be easy to add a command line option to add a transcript length threshold below which transcripts are ignored for analysis (e.g. 300 bases).
Some quantifiers (e.g. Cufflinks) can produce vastly inflated FPKM values for very short transcripts, in which case it is recommended to simply ignore the values produced for such isoforms. Unfortunately piquant converts all FPKM values into TPMs for analysis, and in the case of very high FPKMs, this conversion leads to nearly all TPMs being close to zero - even though most calculated FPKMs are still reasonable.
It will be easy to add a command line option to add a transcript length threshold below which transcripts are ignored for analysis (e.g. 300 bases).