[arvid]

... I have used susceptible、resistance 、susceptible and resistance assebled transcriptomes as reference to compare the number of paired-maping reads、singled-mapping reads、up expressed transcripts.The result of stored in Attach Files.
According to your question(the software and statistics used, the amount and type of replicates, the biological system you are working on, and the way the samples were collected).I haven't used statistics、replicateds yet...

Alright, so don't expect much from your experiment, since you have different strains and no biological replicates... You specify "RPKM>2" in your table, was this log-transformed RPKM values or just raw RPKM? In any case, I really wouldn't recommended using an arbitrary expression-level focused cutoff, and I guess you can't evaluate those thousands of genes anyway. You might want to feed your counts through edgeR or DESeq in R/Bioconductor instead and use the options for experiments without biological replicates (there are good instructions in their respective manuals, or just search SEQanswers). You will not see a lot of differentially expressed transcripts then, but hopefully you can reduce the number of falsely called differentially expressed transcripts. You should afterwards check that the observed differential expression is not due to lower mapping due to polymorphisms in one of the samples, however.
You could try IsoEM or RSEM as well, though I'm not sure how sensitive their respective EM algorithms are to polymorphisms in the underlying mappings. We used RSEM in a similar case to yours (I wasn't involved until a late stage), and were able to identify some qualitatively DE genes (confirmed by RT-qPCR). I can't recommend it again though - replicates are really necessary to get information of high quality out of such an experiment.

[shoegame2001]

I apologize if this question has been addressed elsewhere--I did a little bit of looking but could not find a consensus:

When combining assemblies of different kmer lengths (in my case using Trinity), what is the best method for eliminating redundancy in the resulting combined assembly to make the most comprehensive reference possible? I have tried using cdhit-est, TGICL, and CAP3, but am still observing redundancy when I blast my assembly to itself. For those who have had success with these tools, what options were you using?

Thanks!