Transmembrane Topology Prediction Methods: A Re-assessment and Improvement by a Consensus Method Using a Dataset of Experimentally Characterized Transmembrane Topologies

The presence of signal peptide in the TM sequence has significantly influenced prediction  performance of the ten selected TM topology prediction methods. Unlike in signal peptide, the  presence of transit peptide in the TM sequence has not significantly affected the prediction  performance of the ten selected TM topology prediction methods. In addition, higher prediction  accuracy can be expected if the signal peptide is removed either before or after prediction  analysis. Regarding the question on which of the two-signal peptide treatment is appropriate removed first or later-the choice depends on the TM topology prediction method used, although  statistically, the gain in prediction accuracy is the same. Furthermore, it is imperative that when  performing genome wide analyses of TM protein topology, the issue on signal peptide presence  should be properly addressed. This study also has demonstrated that signal peptide should be  removed if present, which can be done either before or after prediction analysis. Although, at  this point, it is still premature to assert strongly as regards the effect of signal peptide on TM  topology prediction until more experimentally characterized topology data of TM proteins with  signal peptide are available for evaluation.

Bioinformatics, 18  (12): 1562-1566