Abstract:RNA-seq has become the most effective method of analyzing alternative splicing events in different types of cells.The first step of processing data of RNA-seq is to exactly align millions of sequencing fragments against the reference sequence,which is called transcriptome sequence alignment.The existing sequence alignment tools for RNA-seq almost rely on canonical splice site signals,which,to some extent,limits the ability to identify novel splice sites.Therefore,we design a method independent from splice site signals,named RNAMap.It divides the sequencing fragments according to overlapping seeds method and scans the reference sequence via sliding windows with left and right anchors.In this way,splice sites within seeds can be identified.The computational experiments indicate that RNAMap not only reaches a precision of over 95%,but also outperforms the existing softwares in recall rate.