A guideline to proteome‐wide α‐helical membrane protein topology predictions

Abstract

For current state‐of‐the‐art methods, the prediction of correct topology of membrane proteins has been reported to be above 80%. However, this performance has only been observed in small and possibly biased data sets obtained from protein structures or biochemical assays. Here, we test a number of topology predictors on an “unseen” set of proteins of known structure and also on four “genome‐scale” data sets, including one recent large set of experimentally validated human membrane proteins with glycosylated sites. The set of glycosylated proteins is also used to examine the ability of prediction methods to separate membrane from nonmembrane proteins. The results show that methods utilizing multiple sequence alignments are overall superior to methods that do not. The best performance is obtained by TOPCONS, a consensus method that combines several of the other prediction methods. The best methods to distinguish membrane from nonmembrane proteins belong to the “Phobius” group of predictors. We further observe that the reported high accuracies in the smaller benchmark sets are not quite maintained in larger scale benchmarks. Instead, we estimate the performance of the best prediction methods for eukaryotic membrane proteins to be between 60% and 70%. The low agreement between predictions from different methods questions earlier estimates about the global properties of the membrane proteome. Finally, we suggest a pipeline to estimate these properties using a combination of the best predictors that could be applied in large‐scale proteomics studies of membrane proteins.

Conserved Loop Cysteines of Vitamin K Epoxide Reductase Complex Subunit 1-like 1 (VKORC1L1) Are Involved in Its Active Site Regeneration

“…HEK293 Cells-The membrane topology of VKORC1L1 was predicted using ten different topology prediction programs (38). Seven out of the ten programs predict four TMDs in VKORC1L1.…”

In silico evaluation of the influence of the translocon on partitioning of membrane segments

“…A guideline to proteome-wide α-helical membrane protein topology has been published recently [29] giving the opportunity to compare the PMIscale predictions with 18 algorithms on control datasets. We compared PMIscale on two benchmark datasets extracted from this work that permit evaluation of membrane-inserted proteins.…”

“…The first dataset is composed of cytosolic proteins without any signal peptide or TM segment. For this dataset, the PMIscale based predictor predicts 2.8% proteins with at least one TM segment which places it as one of the three best methods, 12% better than the average performance of the evaluated programs in [29]. The second dataset is composed of extracellular proteins that contain a signal peptide but no TM segment.…”

“…To evaluate this point, we used two benchmark datasets. One is composed of 1311 G-protein coupled receptors (GPCRs) extracted from [29]. The particularity of this data set is that the topology of GPCRs is challenging to predict, as several of the TM helices are uncharacteristically hydrophilic.…”

Characterizing the Structure, Function, and Evolution of Human Solute Carrier (SLC) Transporters Using Computational Approaches

“…For example, the support vector machine program MEMSAT-SVM, which was trained on structurally derived topology data and sequence profiles, achieved~90 % accuracy on a set of experimentally determined membrane protein structures (Nugent and Jones 2009). A recent study comprehensively evaluated the performance of topology prediction methods on various sets including a set of 103 newly determined atomic structures (Tsirigos et al 2012). This analysis indicated that the best performance is obtained by TOPCONS (Bernsel et al 2009), a consensus method that combines five other prediction methods (e.g., SCAMPI (Bernsel et al 2008)) and that the best method for distinguishing membrane from non-membrane proteins is the Hidden Markov Model (HMM)-based method Phobius (Kall et al 2005).…”