When dealing with small anchor fraction, the cufflinks would filter out reads before assembly, with the option of "-A/--small-anchor-fraction".
But does cuffdiff do the same filter process?
For example, in my data, anchor of only 2 bases of a read was mapped to gene A. If the tiny part of the read was counted as a fragment by cuffdiff, the FPKM of gene A could be higher than it's real value.
Have anyone met the same problem and how to deal with it?
But does cuffdiff do the same filter process?
For example, in my data, anchor of only 2 bases of a read was mapped to gene A. If the tiny part of the read was counted as a fragment by cuffdiff, the FPKM of gene A could be higher than it's real value.
Have anyone met the same problem and how to deal with it?