Background: Over the last decade,it has been realized that structural variation (SV) is quite abundant in human genome and can exert an influence on major phenotypes.SV is not only the underlying cause of many diseases,but also is a driving force of evolution,leading to phenotypic variations of a trait,ecological adaptation,and speciation.Though the emerging Next Generation Sequencing technology provides high through-put sequence data and quantified gene copy number data,these revolutionary advances still remain powerless in detecting some specific SVs such as micro-inversion.Inversions of DNA segments shorter than 100bp,defined as micro-inversions (MIs),are reported to provide important clues for unveiling disease mechanisms.Unlike the progresses made in MI functional study,our understanding and strategies for detecting and analyzing MIs remain limited.It is notable that reads with potential MIs are usually discarded as "unmapped short reads" by sequence mapping tools,while the existing SV detection tools mainly rely on read mapping.Thus information carried by those reads,which might be important,has been simply ignored.