Skip to main content
NCBI home page
Nucleic Acids Research logoLink to Nucleic Acids Research
. 2022 May 24;50(W1):W29–W35. doi: 10.1093/nar/gkac391

AlignMe: an update of the web server for alignment of membrane protein sequences

René Staritzbichler 1,✉, Emily Yaklich 2, Edoardo Sarti 3, Nikola Ristic 4, Peter W Hildebrand 5,6, Lucy R Forrest 7,✉
PMCID: PMC9252776  PMID: 35609986

Abstract

The AlignMe web server is dedicated to accurately aligning sequences of membrane proteins, a particularly challenging task due to the strong evolutionary divergence and the low compositional complexity of hydrophobic membrane-spanning proteins. AlignMe can create pairwise alignments of either two primary amino acid sequences or two hydropathy profiles. The web server for AlignMe has been continuously available for >10 years, supporting 1000s of users per year. Recent improvements include anchoring, multiple submissions, and structure visualization. Anchoring is the ability to constrain a position in an alignment, which allows expert information about related residues in proteins to be incorporated into an alignment without manual modification. The original web interface to the server limited the user to one alignment per submission, hindering larger scale studies. Now, batches of alignments can be initiated with a single submission. Finally, to provide structural context for the relationship between proteins, sequence similarity can now be mapped onto one or more structures (or structural models) of the proteins being aligned, by links to MutationExplorer, a web-based visualization tool. Together with a refreshed user interface, these features further enhance an important resource in the membrane protein community. The AlignMe web server is freely available at https://www.bioinfo.mpg.de/AlignMe/.

Graphical Abstract

Graphical Abstract.

Graphical Abstract

Open in a new tab

AlignMe Web Server: creating and visualizing pairwise sequence alignments of even distantly-related membrane proteins, by incorporating diverse descriptors, including biochemical data.

INTRODUCTION

Of all proteins, those that span lipid membranes are among the most challenging to compare at the primary sequence level. The reasons are two-fold: first, a membrane-spanning segment requires a high component of hydrophobic amino-acids to match the hydrophobicity of the lipid tails, which reduces the compositional complexity; second, membrane proteins are highly divergent in sequence space (1). These factors are compounded by the fact that membrane proteins have historically been more challenging to examine with structural biology methodologies; the resultant limited coverage of structural space within the Protein DataBank (2) (https://blanco.biomol.uci.edu/mpstruc/) has reduced the availability of data sets for method training and testing (3).

AlignMe (for Alignment of Membrane proteins) is one of only a few methods designed specifically for, and demonstrated to accurately align, membrane protein sequences pairwise (4,5). Other membrane-protein specific methods exist, but most are aimed at multiple-sequence alignments (PRALINE™ (6), TM-Coffee (7), TM-Aligner (8)). Here, we focus on pairwise alignments with the goal of attaining the high accuracy levels (9) required for mechanistic and modeling studies, albeit potentially at the expense of speed. AlignMe is at heart a Needleman-Wunsch global pairwise alignment approach. Its strength is the ability to flexibly incorporate many different types of protein descriptors. Furthermore, the affine gap penalties were trained against sets of structure-based alignments of differing levels of similarity while using a range of different membrane protein-related descriptors (notably hydrophobicity, homologous sequences, secondary structure predictions and transmembrane predictions). This training process resulted in several different modes of operation known as Fast, P, PS or PST (Figure 1), which have increasingly complex inputs and consequently more accurately align sequences with increasing divergence. (In this nomenclature, the P indicates position-specific substitution matrices, S indicates secondary structure and T indicates transmembrane prediction.) In particular, the PST mode, which uses transmembrane predictions from OCTOPUS (10) as one of its inputs, performed well at globally aligning sequences in the midnight zone of sequence similarity (<15%), the range that is least well addressed by all available methods (4). A related but separate feature of AlignMe is the ability to align hydropathy profiles (e.g. obtained from two multiple-sequence alignments), which has facilitated comparisons of complex membrane protein topologies (11–15).

Figure 1.

Figure 1.

Open in a new tab

Input page for pair-wise alignments on the AlignMe web server. The input parameters can be standardized or highly customized. The option to add restraints—or anchors—to the alignment at specific positions with different weights is a recent update to the webserver. The info buttons provide detailed explanations of formatting requirements.

The AlignMe web server (https://www.bioinfo.mpg.de/AlignMe/) was launched over a decade ago as an intuitive interface to the range of functions that AlignMe can provide (16). For simplicity, the input page for aligning family-averaged hydropathy profiles—obtained from two multiple-sequence alignments—is separated from the input page for traditional sequence-based pairwise alignments. In both cases, the web server automatically reports visual representations of the matching of any profile-style descriptors such as secondary structure elements, hydrophobicity, or transmembrane predictions. Moreover, the alignments are provided together with visualizations of matched structural elements (e.g. helices, transmembrane segments) and information such as sequence identities that aid interpretation and modeling applications. The usage of the server has remained relatively consistent throughout this time, with ∼1000 alignments annually.

One of the goals of developing AlignMe was the facilitation of high-quality alignments for use in homology modeling (also known as comparative modeling). In the era of high-reliability structural modeling from machine learning methods such as AlphaFold2 (17) and RosettaFold (18), happily, this application of pairwise alignments is now less frequently needed. Nevertheless, expert-guided sequence alignments are still required for direct comparison of specific proteins at the individual residue level, for guiding biochemical studies of related membrane proteins, including for natively unstructured regions of membrane proteins (19) as well as for more sophisticated structural modeling applications, such as alternate conformations of highly dynamic proteins (20–22), or protein–ligand complexes (23–25). Furthermore, even the most recent machine learning models are not universally accurate in all aspects, meaning that comparison of such models with one another – and with structures and sequences of related proteins – remains an important step. Consequently, web services that provide ready access to such comparison tools will continue to serve an important role.

During the last 10 years, we noticed a few opportunities for enhancements to AlignMe and to the server, which led to the introduction of three new features: anchoring (26); batch processing; and mapping of alignment features onto structures using the web-based server, MutationExplorer. At the same time, we ported the AlignMe software containing these features to GitHub to enhance transparency and reproducibility and to facilitate updates (https://github.com/Lucy-Forrest-Lab/AlignMe). Additionally, we transferred the user manual to GitHub Pages for improved access and simpler navigation (https://lucy-forrest-lab.github.io/AlignMe/). In this article, we describe some of the new features in the context of the web server, which the user will also find has been modernized and refreshed since the original publication (16).

ENFORCING MATCHING OF KNOWN RELATED POSITIONS USING ANCHORS

When comparing protein structures with very low sequence identities and/or with many membrane-spanning segments, many global alignment methods fail to even match up the transmembrane elements correctly across the entire span of the protein. In such cases, minimum amounts of information from biochemical, biophysical, or genetic data, such as the identification of residues or short motifs involved in substrate binding, lipid interactions, disease states or folding, can be transformative in guiding a more accurate global alignment. However, incorporation of such information has typically relied on the expert manually modifying the alignment, i.e. shifting segments of the alignment so that key positions are matched (e.g. (27)). Such manipulations are undesirable since they arbitrarily remove information provided by the alignment algorithm in the remainder of the alignment, especially at the regions immediately adjacent to the matched positions. Other alignment tools, such as SALIGN, can constrain individual residues (28), while DIALIGN (29,30) and COBALT (31) modify large-scale alignments for genomic studies based on fragments or domain information. A few methods can constrain sequence motifs, as in MA-PRALINE (32), ConBind (33) and FMALIGN (34). To automate the incorporation of this level of information in membrane protein alignments we added a similar feature, called anchoring, into AlignMe v1.2.2 (26).

To add anchors to an alignment, the user provides the amino acid positions to be matched—one for each of the two input sequences. A constraint strength factor, or weight, is also required. However, we recommend the use of a weight that universally imposes the matching of the two residues; specifically, weight values >10 will typically be sufficient to match the two positions (26), no matter the complexity or sequence similarity of the sequences. Accordingly, the default value for anchor weights is 1000. However, by gradually increasing the weight in the range of 0.1–100 and assessing the impact on the alignment (i.e. whether the two positions are matched, and what is the effect on the local region on the alignment), the optimum value can be identified for a given pair of sequences. Note that the number of anchors is not limited to one pair of matched positions. Indeed, an anticipated use of anchors will be in attaining a full-length alignment of two sequences for which large fragments have previously been obtained using structure-based alignments of structural repeats for the EncoMPASS database (35). Obviously, using more than one anchor creates the potential for mutually incompatibility. In such a situation, computing alignments with each of the anchors separately should allow the user to assess their relative impacts and interdependence (e.g. in the context of an available structure, using the molecular viewer described below).

In the context of the web server, the anchor function is now provided in the main pair-wise alignment web page as an additional input section. Here, the residue numbers in each of the two sequences must be entered, along with the weight or strength (Figure 1). Where a long list of anchors is required, a file may be uploaded containing all the desired matching pairs.

After submission, the alignments are computed while enforcing these anchors. AlignMe does so at the level of the dynamic programming similarity matrix, which under normal conditions is populated by scores reflecting (i) the matching of the usual input descriptors, (ii) gap penalties and (iii) the scores of neighboring cells. To impose an anchor, a high repulsive score is added to regions of the grid surrounding the matched pair. This effectively narrows the universe of possible paths that the alignment can take during the traceback procedure. Only when the anchor strength is set to a value (∼1) close to that of typical matching scores would the traceback through the matrix take a path that does not include the exact specified positions.

Anchored positions in AlignMe alignments are clearly labeled in the output (Figure 2). Repeating the alignment without providing anchor(s) allows a straightforward analysis of their effect(s).

Figure 2.

Figure 2.

Open in a new tab

Example pairwise alignment obtained using the AlignMe web server either without (upper) or with (lower) anchoring. The letter ‘a’ indicates the anchored positions on the alignment. Note that when the weight of the anchor is too small to successfully impose the matching of the two residues, then there will be no letter ‘a’ shown. Identical positions are highlighted with ‘*’. The sequences are sarcoendoplasmic reticulum calcium transport ATPase (SERCA) regulators: mouse endoregulin (mELN) and Drosophila sarcolamban isoform b (dsCLB).

RUNNING BATCHES OF INDIVIDUAL PAIRWISE ALIGNMENTS

A common situation faced by users of the AlignMe web server was the need to run multiple pairwise alignments for different sequences, for example, when comparing many different templates for a homology model. Previously, when using the web server, the user would be required to initiate each pairwise alignment individually through the main input page. Although the AlignMe software can be easily accessed via GitHub for local installation in multi-use cases, usage of all its functions requires several dependencies: BLAST, PSIPRED and OCTOPUS. As some of these are complex codes, their installation imposes significant time and expertise requirements, while limiting the range of usable operating systems. Therefore, to streamline access to AlignMe and to make large-scale calculations available to all users, we introduced an option to run batches of alignments on the server with a single set of input parameters.

Batch processing on the AlignMe web server is trivial to access. Specifically, rather than providing a single sequence in each of the two input boxes (Figure 1) – or in the files for upload – the user provides all the sequences to be aligned. The format of the input is explained in information pop-ups. For the provided sets of sequences, every sequence in the first input will be aligned with every sequence in the second input, up to a maximum of 1000 total alignments per submission. Every alignment will be pairwise and will use the same set of parameters. (We note that this is not the same as a multiple-sequence alignment, which is a feature that we are considering introducing in future versions of AlignMe.) As for all calculations on the web server, the user can provide an e-mail address if they wish to be alerted when all alignments have been completed.

The results page for AlignMe can contain a lot of information, even for a single pair of sequences: it shows the alignment along with profile plots (Figure 3) illustrating the matching of secondary structure predictions, hydrophobicity, or transmembrane predictions, as appropriate, plus a numerical representation of the output. All this information can also be downloaded. When processing a batch of alignments, clearly a very large number of possible results might be obtained. Therefore only a single output is provided on the results page—using the usual formatting—for the alignment between the first sequence in each of the two inputs. The header of the results page explains that multiple different alignments were processed and the complete set of results, including plots for all pairs, are available via download links (Figure 3).

Figure 3.

Figure 3.

Open in a new tab

Results page obtained when running a batch of alignments simultaneously on the AlignMe web server. (A) The top of the page reports the number of computed alignments, and detailed information for the alignment of the first pair of sequences is shown, including the alignment itself. The link to the MutationExplorer visualization tool is also visible here. (B) Further down the page, plots of profile-type alignments are shown, along with links for downloading the collection of results for all pairs of sequences. The example sequences are mouse endoregulin (mELN) and mouse another-regulin (mALN).

A combination of the two new features, i.e. batches of anchored alignments, is also possible. In this case, the user simply needs to specify the sequence number in addition to the residue number in the input file listing the requested anchors. The results page and downloadable results files highlight any applied anchors on the corresponding sequence alignments (Figure 2).

We note that currently, to align batches of sequences using profile representations, the web server can automatically compute profiles of the following three types: (i) hydropathy profiles based on hydrophobicity scales; (ii) PSIPRED secondary structure predictions or (iii) OCTOPUS transmembrane helix predictions. User-generated profiles are not currently compatible with batches of alignments, and so for the time being, the user is encouraged to install the software locally if they require that specific advanced functionality.

VISUALIZATION OF ALIGNMENTS IN A STRUCTURAL CONTEXT

Alignments of primary sequences, especially long ones, can be challenging to interpret, since they are simplifications of complex three-dimensional molecules. The plots of aligned secondary structure or transmembrane probabilities provided by the AlignMe web server help envisage those 3D structures since they represent the relationship between the proteins in different regions of the alignment. Nevertheless, a feature to relate the matched (or mismatched) segments directly in the context of an atomic-resolution structure or a structural model would provide an entirely new level of insight. For example, if a structure is known for one of the two proteins being aligned, the location of sequence differences in structural space informs on their functional implications.

To this end, the AlignMe server now provides the option to map the sequence similarities identified in the AlignMe alignment onto one or more structures. Specifically, every pairwise alignment results webpage now includes a link to a molecular structure viewer at the MutationExplorer web server (http://proteinformatics.uni-leipzig.de/mutation_explorer/; Figure 3). With this option, the user can upload a structural coordinate file that corresponds to a sequence in the alignment, or, if available in the Protein DataBank (2), provide the associated four-character identifier for automatic retrieval. A second coordinate file or PDB identifier corresponding to the other sequence can also be given. The MutationExplorer viewer will automatically relate the alignment to the structure; that is, it will find the first chain in the structure file that maps exactly onto a sequence in the alignment. The structure(s) will then be loaded into a customized molecular viewer for quick visualization. Importantly, the coloring of the matched chain(s) indicates the identical, similar, and gapped residues in the AlignMe alignment (Figure 4). If two structures were provided, they will both be colored in the same way, and will also be superposed according to the AlignMe alignment.

Figure 4.

Figure 4.

Open in a new tab

Visual representation of AlignMe results mapped onto the structure corresponding to one of the two sequences. AlignMe was used to align the sequences of cytochrome b6 (PDB entry 2ZT9) and Ni/Fe-hydrogenase 1 cytochrome b (PDB entry 4GD3 chain A) using the PST mode. After the alignment was obtained, a set of coordinates was uploaded and sent together with the alignment to the MutationExplorer viewer at http://proteinformatics.uni-leipzig.de/mutation_explorer for quick visualization. According to the AlignMe alignment, the residues in chain A of 2ZT9 are identical (red), similar (yellow), or different (light blue) from those of cytochrome b. Dark blue coloring indicates unaligned regions, i.e. gaps or chains in the complex other than the aligned protein chain. This result was obtained using the ‘Generate example input’ option on the pair-wise alignment page shown in Figure 1—see text for details.

Trying out the structure viewer is made easy by using the ‘Generate example input’ option on the main pair-wise alignment page (Figure 1). Then, on the results page, after clicking ‘MutationExplorer’, the PDB identifier of one or both sequences (4GD3 and 2ZT9) should be entered into the text boxes, before clicking submit.

We note that this feature is available for any of the alignments computed as part of a batch, as the user can select the alignment name prior to selecting the protein structure.

For the expert viewer, in-depth visualization of the same structure(s) is readily accessible from this customized quick MutationExplorer viewer, through a hyperlink. This full-screen version provides the full capability of the molecular viewer. We anticipate that this sequence-structure mapping feature will enable a rapid interpretation of alignment information for membrane proteins for which structural information or prediction is available. For example, it will aid rapid comparison of results from threading of a homology model or from application of anchors to a given position in an alignment.

CONCLUSIONS

The AlignMe web server has been a useful resource for the membrane protein research community for over a decade. A high level of customizability means that AlignMe could also, in principle, be readily applied to obtain full-length alignments of non-membrane proteins, e.g. low complexity unstructured proteins. Thus, it may be of interest to pursue systematic studies of AlignMe to such applications. In addition, we plan to extend the AlignMe functionality to generate multiple alignments using its high-accuracy pairwise alignments as a foundation. In the meantime, the additional updates described here enhance the accessibility, functionality, and usability of the web server and provide a framework for such studies in the future.

ACKNOWLEDGEMENTS

The web server has been continually maintained by the HPC application support team at the Max Planck Gesellschaft – we thank Drs Markus Rampp and Mykola Petrov for their efforts. We also thank Drs Leone and Hellsberg from NINDS, and Ms. Park from NHLBI for testing of the MutationExplorer functions.

Contributor Information

René Staritzbichler, University of Leipzig, Institute of Medical Physics and Biophysics, Härtelstr. 16-18, 04107 Leipzig, Germany.

Emily Yaklich, Computational Structural Biology Section, National Institutes of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892, USA.

Edoardo Sarti, Algorithms, Biology, Structure Unit Inria Sophia Antipolis – Méditerranée, 06902 Valbonne, France.

Nikola Ristic, University of Leipzig, Institute of Medical Physics and Biophysics, Härtelstr. 16-18, 04107 Leipzig, Germany.

Peter W Hildebrand, University of Leipzig, Institute of Medical Physics and Biophysics, Härtelstr. 16-18, 04107 Leipzig, Germany; Charité –Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin, Institute of Medical Physics and Biophysics, 10117 Berlin, Germany.

Lucy R Forrest, Computational Structural Biology Section, National Institutes of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892, USA.

FUNDING

Max Planck Gesellschaft; Division of Intramural Research of the NIH; National Institute of Neurological Disorders and Stroke; DFG (German Research Foundation) [CRC 1423, project number 421152132 subproject Z04 to P.W.H.]. Funding for open access charge: National Institute of Neurological Disorders and Stroke.

Conflict of interest statement. None declared.

REFERENCES

  • 1. Sojo V., Dessimoz C., Pomiankowski A., Lane N.. Membrane proteins are dramatically less conserved than water-soluble proteins across the tree of life. Mol. Biol. Evol. 2016; 33:2874–2884. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Berman H.M., Westbrook J., Feng Z., Gilliland G., Bhat T.N., Weissig H., Shindyalov I.N., Bourne P.E.. The protein data bank. Nucleic Acids Res. 2000; 28:235–242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Bahr A., Thompson J.D., Thierry J.C., Poch O.. BAliBASE (benchmark alignment dataBASE): enhancements for repeats, transmembrane sequences and circular permutations. Nucleic Acids Res. 2001; 29:323–326. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Stamm M., Staritzbichler R., Khafizov K., Forrest L.R.. Alignment of helical membrane protein sequences using AlignMe. PLoS One. 2013; 8:e57731. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Hill J.R., Deane C.M.. MP-T: improving membrane protein alignment for structure prediction. Bioinformatics. 2013; 29:54–61. [DOI] [PubMed] [Google Scholar]
  • 6. Pirovano W., Feenstra K.A., Heringa J.. PRALINE (TM): a strategy for improved multiple alignment of transmembrane proteins. Bioinformatics. 2008; 24:492–497. [DOI] [PubMed] [Google Scholar]
  • 7. Floden E.W., Tommaso P.D., Chatzou M., Magis C., Notredame C., Chang J.M.. PSI/TM-Coffee: a web server for fast and accurate multiple sequence alignments of regular and transmembrane proteins using homology extension on reduced databases. Nucleic Acids Res. 2016; 44:W339–W343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Bhat B., Ganai N.A., Andrabi S.M., Shah R.A., Singh A.. TM-Aligner: multiple sequence alignment tool for transmembrane proteins with reduced time and improved accuracy. Scientific Rep. 2017; 7:12543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Wang Y., Wu H., Cai Y.. A benchmark study of sequence alignment methods for protein clustering. BMC Bioinformatics. 2018; 19:529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Viklund H., Elofsson A.. OCTOPUS: improving topology prediction by two-track ANN-based preference scores and an extended topological grammar. Bioinformatics. 2008; 24:1662–1668. [DOI] [PubMed] [Google Scholar]
  • 11. Khafizov K., Staritzbichler R., Stamm M., Forrest L.R.. A study of the evolution of inverted-topology repeats from LeuT-fold transporters using AlignMe. Biochemistry. 2010; 49:10702–10713. [DOI] [PubMed] [Google Scholar]
  • 12. Fenollar-Ferrer C., Patti M., Knopfel T., Werner A., Forster I.C., Forrest L.R.. Structural fold and binding sites of the human Na(+)-phosphate cotransporter NaPi-II. Biophys. J. 2014; 106:1268–1279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Lara P., Öjemalm K., Reithinger J., Holgado A., Maojun Y., Hammed A., Mattle D., Kim H., Nilsson I.. Refined topology model of the STT3/Stt3 protein subunit of the oligosaccharyltransferase complex. J. Biol. Chem. 2017; 292:11349–11360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Ratcliffe S., Jugdaohsingh R., Vivancos J., Marron A., Deshmukh R., Ma J.F., Mitani-Ueno N., Robertson J., Wills J., Boekschoten M.V.et al.. Identification of a mammalian silicon transporter. Am. J. Physiol. Cell Physiol. 2017; 312:C550–C561. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Ballesteros A., Fenollar-Ferrer C., Swartz K.J.. Structural relationship between the putative hair cell mechanotransduction channel TMC1 and TMEM16 proteins. Elife. 2018; 7:e38433. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Stamm M., Staritzbichler R., Khafizov K., Forrest L.R.. AlignMe–a membrane protein sequence alignment web server. Nucleic Acids Res. 2014; 42:W246–W251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Jumper J., Evans R., Pritzel A., Green T., Figurnov M., Ronneberger O., Tunyasuvunakool K., Bates R., Zidek A., Potapenko A.et al.. Applying and improving AlphaFold at CASP14. Proteins. 2021; 89:1711–1721. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Baek M., DiMaio F., Anishchenko I., Dauparas J., Ovchinnikov S., Lee G.R., Wang J., Cong Q., Kinch L.N., Schaeffer R.D.et al.. Accurate prediction of protein structures and interactions using a three-track neural network. Science. 2021; 373:871–876. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Atsmon-Raz Y., Miller Y.. Non-Amyloid-β component of human α-Synuclein oligomers induces formation of new Aβ oligomers: insight into the mechanisms that link Parkinson's and Alzheimer's diseases. ACS Chem. Neurosci. 2016; 7:46–55. [DOI] [PubMed] [Google Scholar]
  • 20. Forrest L.R., Zhang Y.W., Jacobs M.T., Gesmonde J., Xie L., Honig B.H., Rudnick G.. Mechanism for alternating access in neurotransmitter transporters. Proc. Natl. Acad. Sci. U.S.A. 2008; 105:10338–10343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Kowalczyk L., Ratera M., Paladino A., Bartoccioni P., Errasti-Murugarren E., Valencia E., Portella G., Bial S., Zorzano A., Fita I.et al.. Molecular basis of substrate-induced permeation by an amino acid antiporter. Proc. Natl. Acad. Sci. U.S.A. 2011; 108:3935–3940. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Liao J., Li H., Zeng W., Sauer D.B., Belmares R., Jiang Y.. Structural insight into the ion-exchange mechanism of the sodium/calcium exchanger. Science. 2012; 335:686–690. [DOI] [PubMed] [Google Scholar]
  • 23. Cavasotto C.N., Orry A.J., Murgolo N.J., Czarniecki M.F., Kocsi S.A., Hawes B.E., O’Neill K.A., Hine H., Burton M.S., Voigt J.H.et al.. Discovery of novel chemotypes to a G-protein-coupled receptor through ligand-steered homology modeling and structure-based virtual screening. J. Med. Chem. 2008; 51:581–588. [DOI] [PubMed] [Google Scholar]
  • 24. Schlessinger A., Khuri N., Giacomini K.M., Sali A.. Molecular modeling and ligand docking for solute carrier (SLC) transporters. Curr. Top Med. Chem. 2013; 13:843–856. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Andersen J., Ladefoged L.K., Wang D., Kristensen T.N., Bang-Andersen B., Kristensen A.S., Schiott B., Stromgaard K.. Binding of the multimodal antidepressant drug vortioxetine to the human serotonin transporter. ACS Chem. Neurosci. 2015; 6:1892–1900. [DOI] [PubMed] [Google Scholar]
  • 26. Staritzbichler R., Sarti E., Yaklich E., Aleksandrova A.A., Stamm M., Khafizov K., Forrest L.R.. Refining pairwise sequence alignments of membrane proteins by the incorporation of anchors. PLoS One. 2021; 16:e0239881. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Nim S., Lobato L.G., Moreno A., Chaptal V., Rawal M.K., Falson P., Prasad R.. Atomic modelling and systematic mutagenesis identify residues in multiple drug binding sites that are essential for drug resistance in the major Candida transporter Cdr1. Biochim. Biophys. Acta - Biomembr. 2016; 1858:2858–2870. [DOI] [PubMed] [Google Scholar]
  • 28. Marti-Renom M.A., Madhusudhan M.S., Sali A.. Alignment of protein sequences by their profiles. Protein Sci. 2004; 13:1071–1087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Morgenstern B., Werner N., Prohaska S.J., Steinkamp R., Schneider I., Subramanian A.R., Stadler P.F., Weyer-Menkhoff J.. Multiple sequence alignment with user-defined constraints at GOBICS. Bioinformatics. 2005; 21:1271–1273. [DOI] [PubMed] [Google Scholar]
  • 30. Morgenstern B., Prohaska S.J., Pohler D., Stadler P.F.. Multiple sequence alignment with user-defined anchor points. Algorithms Mol. Biol. 2006; 1:6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Papadopoulos J.S., Agarwala R.. COBALT: constraint-based alignment tool for multiple protein sequences. Bioinformatics. 2007; 23:1073–1079. [DOI] [PubMed] [Google Scholar]
  • 32. Dijkstra M., Bawono P., Abeln S., Feenstra K.A., Fokkink W., Heringa J.. Motif-Aware PRALINE: improving the alignment of motif regions. Plos Comput. Biol. 2018; 14:e1006547. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Lelieveld S.H., Schutte J., Dijkstra M.J., Bawono P., Kinston S.J., Gottgens B., Heringa J., Bonzanni N.. ConBind: motif-aware cross-species alignment for the identification of functional transcription factor binding sites. Nucleic Acids Res. 2016; 44:e72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Chakrabarti S., Bhardwaj N., Anand P.A., Sowdhamini R.. Improvement of alignment accuracy utilizing sequentially conserved motifs. BMC Bioinformatics. 2004; 5:167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Sarti E., Aleksandrova A.A., Ganta S.K., Yavatkar A.S., Forrest L.R.. EncoMPASS: an online database for analyzing structure and symmetry in membrane proteins. Nucleic Acids Res. 2019; 47:D315–D321. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press

RESOURCES