Trinity would be a good start to assemble a transcriptome. Beware you will need a server with plenty of RAM to run this. For annotation, is it prokaryote or eukaryote? The MAKER pipeline is known for eukaryotes, but there are plenty of choices.

I agree with Johnathan. Trinity is always my first choice. If the dataset is really huge and your RAM is not enough, run the digital normalization (diginorm) script that is in the Trinity package. It will discard reads based on its constituint k-mer frequencies in the whole dataset. This way, you reduce the coverage of highly covered transcripts (and, as a bonus, the errors they contain). The diginorm procedure will not discard reads whose kmers frequencies are bellow the threshold, so low-coverage transcripts are preserved. The RAM requirements after diginorm are significantly lower. Surprisingly, almost all assemblies benefit from diginorm.

Thanks Jonathan and Gustavo. I shall use trinity and let u know if I have some problem.

You can try to use Annocript, a pipeline to perform annotation of assembled transcriptomes

http://www.ncbi.nlm.nih.gov/pubmed/25701574

Available at: https://github.com/frankMusacchia/Annocript