Real Or Ghost Tm Domains

Question

[cdsouthan]

This is a Ciona savignyi Ur-BACE ORF from Ensembl

MVKFLWCLWLGVFIVCSGARIRLKRAATDNLHGTTITGYYVTVKLGNPPQELNVIVDTGSSNFAVAGGPNPVISNFYNKSSSNTSVDTGVRSVKVDYTEGWWMGDVVSDVLSVPSAQLDTSVRVPVADITNSNKFFVNGSNWVGILGLAYSDLVLPKGNGLKSVMHEITHQTSTPDLLSMQLCSTSLQDATYGALLIGEIDLSLAAGPLYWTPIVKQWYYDIIVSGLKIGDKVVDIDCSDINYDRTIVDSGTTNLRFPQKVYDIILPIIKASVVRFYFDQDFYDGKTMFCTTDPAELYGEFPNITIYLPSTNENQTIELTIHSASIPQFTTMLDASKQSCYKFAIFPSAIGTVLGTVLMEEYYVVFNRINSSIGFALSACHKGSTIQVINKWINSTNCRYIPPTKPITYAAYVVLALFCICLLPLCFLLAFNYFRSKSTPCRPRESYDQLDDHHEVSH

When I run this through TM prediction at http://www.cbs.dtu.dk/services/TMHMM/ , in addtion to the expected score at the signal peptide and the C-terminal TM I see another high score segment at 339-359.

Could this "cryptic TM" be real and significant in protein function/location terms ?

As we know these TMs don't show high sequence conservation so homology comparison is not that useful

[sarahhunter]

Hi Chris

Have you tried running it through other TM predictors, like Phobius? The signal is much lower there...

Sarah

[cdsouthan]

Your right, thanks. OK so I'll chalk up this particular cryptic TMHMM as an artefact, but this post can ensure it persists in the collective memory bank, just in case it comes up again.

Answer 1

[jackuser1979]

Whenever you do bioinformatics predictions, it always better to get consensus view by using different predictors and decide if all the predictors give the same TM prediction. See the list of TM predictors