RSEM: differential expression of specific enzymes within transcriptome

[leontovyc_roman]

Hi
I'm interested in specific proteolytic enzymes in transcriptomic data. I chose 3 different transcripts from de novo assembly and run RSEM analysis with pair-end data reads.
First question: Is it Ok to run RSEM on such small set of reference sequences?

Second question: How to deal with different length of reference sequences? In my case I had 980, 247 and 929 nucleotides. I run two analysis, one with original length of reference sequences, and second with trimmed sequences on 247 nucleotides for all sequences and I got different results.

thanks in advance.

Roman