Analysis of slipped sequences in ESTs Projects

Slippage is an important sequencing problem that can occur in EST projects. However, there are very few studies about it. In this work we propose three new methods to detect slippage artifacts: “Arithmetic Mean Method”, “Geometric Mean Method”, and “Echo Coverage Method”. Each method is simple and has two different strategies for processing sequences: suffix and subsequence. Using the 291689 EST sequences produced in the SUCEST project [9], we performed comparative tests between the proposed methods and Telles and Silva Method [8]. The subsequence strategy is better than the suffix strategy because it is not anchored at the end of the sequence, so it is more flexible to find slippage at the beginning of the EST. Comparing with the Telles and Silva Method, the advantage of our methods is that they do not discard the majority of the sequences marked as slippage, but, instead of it, only remove the slipped artifact from the sequence. The tests indicate that the “Echo Coverage Method” with subsequence strategy has the best compromise between slippage detection and calibration easiness.