The RNA-Seq is a recently developed method to profile transcriptome including mRNAs and small RNAs. This revolutionary tool for transcriptomics, RNA-seq, altered the view of the extent and complexity of transcriptomes. Using RNA-seq, researchers can efficiently measure transcriptome data experimentally, they can get information such as gene fusions, differential expression of alleles of genes, or existences of small RNA and post-transcriptional mutations. To harness this powerful tool we need new methods and tools to overcome the chanllenge which come with RNA-seq. In this thesis, we proposed methods and systems which integrate biological knowledge and algorithmic techniques to improve the accuracy and efficiency of transcriptome profiling.