The gene structure was optimized according to the distribution of the
reads, information of paired-end and the annotation of reference gene. We can
get the distribution of reads in the genome by aligning the continuous and
overlap reads form a Transcription Active Region (TAR). According to
paired-end data, we can connect the different TARs to form a potential gene
model. We can compare the gene model with the existing gene annotated to
extend the gene 5'and 3' end