Detecting the impact of sequencing errors on SAGE data

SAGE data are obtained by sequencing short DNA tags. Due to the mistakes in DNA sequencing, SAGE data contain errors. We propose a new approach to ide ntify tags whose abundance is biased by sequencing errors. This approach is based on a concept of neighbourhood: abundant tags can contaminate tags wh ose sequence is very close. The application of our approach reveals that mo derately abundant tags can be generated by sequencing errors uniquely. It a lso allows for detecting correct rare tags.