Affymex: a bioinformatics online tool for optimizing abundant protein depletion and mass spectrometry analysis.
Kosi Gramatikoff,Xiangming Fang,Wei-Wei Zhang
DOI: https://doi.org/10.2144/05386SU05
2005-01-01
BioTechniques
Abstract:BioTechniquesVol. 38, No. 6S Featured ArticleOpen AccessAffymex: a bioinformatics online tool for optimizing abundant protein depletion and mass spectrometry analysisKosi Gramatikoff, Xiangming Fang & Wei-Wei ZhangKosi Gramatikoff*Address correspondence to: Kosi Gramatikoff GenWay Biotech 10130 Sorrento Valley Road Suite C San Diego, CA 92121, USA. e-mail: E-mail Address: kosi@genwaybio.comGenWay Biotech, San Diego, CA, USASearch for more papers by this author, Xiangming FangGenWay Biotech, San Diego, CA, USASearch for more papers by this author & Wei-Wei ZhangGenWay Biotech, San Diego, CA, USASearch for more papers by this authorPublished Online:30 May 2018https://doi.org/10.2144/05386SU05AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinkedInRedditEmail Specific removal of high-abundant proteins from a proteomic sample can improve the ability of mass spectrometry (MS) to detect low-abundant proteins and enhance biomarker discovery (1,2). Immunoaffinity separation of proteins is one of the best methods for this purpose (3–5). Peptide mass fingerprints have been extensively used for MS to identify proteins and develop proteomic profiles (6–8). However, there is no method to date to predict how a particular high-abundant protein interferes with the MS analysis of specific low-abundant proteins or biomarkers. In order to rationally guide depletion of the high-abundant proteins for MS analysis, we applied peptide mass fingerprints and developed an online program (Affymex) for analyzing molecular interference between high-abundant and low-abundant proteins. Affymex generates isobaric charts by quantifying the degree of overlap of peptide mass fingerprints derived from an interfering high-abundant protein with that from a target low-abundant protein. This degree of overlap is extrapolated by a single integer called mass match value (MMV), which measures how much a particular high-abundant protein would interfere with specific low-abundant proteins during MS analysis. As illustrated in Figure 1, the closer two peptide masses, the higher the MMV generated, and the greater the predicted interference between them. MMV is a central parameter that can be correlated to a variety of other parameters such as molecular weight, isoelectric point, concentration, posttranslational modifications of intact proteins, and/or their peptide mass fingerprint. The data can be plotted in 3-D for visualization of the proteomic landscape. The algorithm was written in C-sharp computer language, with a graphical module written in VB.NET hosted on a Microsoft®.NET web server. The database architecture on the web server was designed to include SwissProt protein sequence database (www.expasy.org) as a primary source of every target protein and an internal database as a secondary source for the predigested high-abundant proteins or other potentially interfering proteins.Figure 1. Principle of isobar generation and assignment of mass match value (MMV).Peptide monoisotopic masses [M+H]+ derived from interleukin-6 (IL-6; a, b, and c, orange bars) and α-2-macroglobulin (A2MG; a, b, and c, blue bars) were generated by PeptideMass (SwissProt) via trypsin digestion and were superimposed to calculate MMV (A, B, C, pink bars).To predict the degree of high-abundant protein interference with low-abundant protein detection, the following algorithm was implemented. Affymex calculates molecular mass proximity of peptides derived from interfering high-abundant protein (iMM) to the molecular mass of the peptides derived from that of a target low-abundant protein (tMM). The degree of overlap of an iMM to a tMM is visualized by isobar (MMV = f[iMM∼tMM]) (Figure 1). The isobars are presented graphically as a function of the isotopic peptide mass of the target protein (Figure 1, x-axis). The closer the iMM to tMM, the higher the MMV (MMVmax [iMM = tMM]). Every isobar is indicated according to its source of matching high-abundant protein and presents the real value of tMM. For example, what would be the degree of overlap of the iMM derived after trypsin digest of a high-abundant protein [α-2-macroglobulin (A2MG)] with the tMM derived from a protein of interest, interleukin-6 (IL-6)? Figure 1 and 2 provide examples of how Affymex was applied to answer this question. When the A2MG-derived peptide with an iMM = 765.4141 (Figure 1b, blue bar) is closed to a peptide from IL-6 (tMM = 765.4002, Figure 1b, orange bar), the absolute difference is ∣0.0139∣. This results in a MMV = 72 assigned by Affymex (Figure 1B, pink bar). The resulted isobar with MMV = 72 on the y-axis, is positioned on the x-axis corresponding to iMM = 765.4141. The exact MMV numerical value can be seen in the browser by rolling over the bar with your cursor. Both high-abundant proteins and target proteins can be digested in variety of methods as enlisted by PeptideMass (www.expasy.org/tools/peptide-mass.html). Interfering abundant proteins can be analyzed by two different modes: (i) grouped high-abundant proteins that can be removed together by the commercially available kits and (ii) 12 individual high-abundant proteins in plasma that can be selected and removed by IgY microbeads (GenWay Biotech, San Diego, CA, USA) (4,5).Using the online tool, we have analyzed 675 SwissProt entries from the human plasma proteome (9). A batch procedure was executed to process simultaneously multiple entries for target proteins (tMM). Out of the 675 sequences, 37 protein entries showed no fragmentation with trypsin and therefore were omitted from the further analysis and presentation. Special attention was given to low molecular weight proteins (up to 20 kDa), since many biomarkers are found in this range. The molecular weight of 638 analyzed proteins of interest, their monoisotopic mass, and MMV data were assigned to the x, y, and z coordinate system and exported to SigmaPlot-9 (www.systat.com/products/SigmaPlot). To visualize the data in 3-D, a smoothing technique was applied using the weighted average of z-values at neighboring points using the Cauchy density function (Figure 3A).The obtained 3-D landscape at low molecular weight range of the plasma proteome shows the monoisotopic mass and molecular weight areas that could be affected by the 12 high-abundant proteins. The MMVmax from the input table were numerically identified, and a few known biomarkers were marked on the 3-D chart around the surface with black arrows. IL-10 appears to be a more affected biomarker as compared to IL-6, its four peptides overlapped by the peptides of two high-abundant proteins, albumin and haptoglobin (Figure 3A). If a multiple target approach for MS analysis of few biomarkers is taken, Affymex can direct and optimize an immunoaffinity depletion scheme. For instance, as can be seen from Figure 3B, depletion of albumin can improve the analysis of IL-10, pancreatitis-associated protein 1, and hepatocyte growth factor activator, but would not bring improvement for the rest of the marked proteins such as IL-18 binding protein, IL-10, S-100 protein chain, neuropeptide Y, and parathyroid hormone.In summary, our data demonstrate that Affymex is a powerful and easily accessible online tool for optimizing specific depletion schemes of abundant proteins before MS analysis. It will evolve to analyze the cross-influence of all possible peptides derived from a particular proteome. As a consequence, very accurate affinity depletion schemes will be designed in silico to remove not only high-abundant proteins that interfere with MS analysis, but also all possible proteins for which we have no current knowledge or prediction. To enhance MS analysis of proteins and proteomic profiling, further principal improvements of the program will be made. Affymex 2-D data as shown in Figure 2, top panel, can be extended into 3-D space by adding a third axis for the display of parameters additional to molecular weight of target proteins (Figure 3), such as peptide isoelectric point, frequency of detection, peptide concentration, and posttranslation modifications. By adding alternative dimensions to the original Affymex 2-D data, it is possible to observe global landscapes of protein data distribution and clusters. Affymex can also be further developed into a clinical diagnostic program.Figure 2. Example of 2-D output of interleukin-6 (IL-6) isobars.High-abundant proteins (HAP; list on the bottom right, as annotated by SwissProt) were virtually digested by trypsin and resulted in 577 peptide masses (step 1), compared to a target protein (IL-6) digested with trypsin (step 2) for matches and nearest neighbors isolated with assignment of mass match values (MMV; step 3, top panel on the left).Figure 3. 3-D visualization of the interference potential of 12 high-abundant proteins over the low molecular weight plasma proteome.(A) The protein names in large bold type correspond to those biomarkers that produce peptide mass fingerprints colliding with the peptide mass fingerprints of 12 high-abundant proteins. The protein names in small type correspond to the source high-abundant proteins that would interfere with the mass spectrometry (MS) analysis of the protein of interest in bold. (B) Extracted information about biomarkers affected by albumin depletion (ALB+) and not affected by albumin depletion (ALB-). MMV, mass match value.AcknowledgmentsAffymex is available at the TOOLS section on the web site main menu of GenWay Biotech (www.genwaybio.com).Competing Interests StatementThe authors declare that they are seeking a patent on the method and related applications described.References1. Anderson, N.L. and N.G. Anderson. 2002. The human plasma proteome history, character, and diagnostic prospects. Mol. Cell Proteomics 1:845–867.Crossref, Medline, CAS, Google Scholar2. Hinerfeld, D., D. Innamorati, J. Pirro, and S.W. Tam. 2004. Serum/plasma depletion with chicken immunoglobulin Y antibodies for proteomic analysis from multiple mammalian species. J. Biomol. Tech. 3:184–190.Google Scholar3. Fang, X., K.W. Curran, L. Huang, W. Xiao, W. Strauss, G. Harvie, J. Feitelson, and W.-W. Zhang. 2003. Polyclonal gene-specific IgY antibodies for proteomics and abundant plasma protein depletion, p. 222–245. In J. Reiner, K. Zhao, S.-H. Chen, and M. Guo (Eds.), Frontiers of Biotechnology and Pharmaceuticals, Vol. 4. Science Press USA, Monmouth Junction, NJ.Google Scholar4. Hinerfeld, D., D. Innamorati, J. Pirro, and S.W. Tam. 2004. Serum/plasma depletion with chicken immunoglobulin Y antibodies for proteomic analysis from multiple mammalian species. J. Biomol. Tech. 15:184–190.Medline, Google Scholar5. Fang, X, L Huang, J Feitelson, and W-W Zhang. 2004. Affinity separation: divide and conquer the proteome. Drug Discovery Today: Technologies 1:141–148.Crossref, Medline, CAS, Google Scholar6. Gevaert, K. and J. Vandekerckhove. 2000. Protein identification methods in proteomics. Electrophoresis 21:1145–1154.Crossref, Medline, CAS, Google Scholar7. Samuelsson, J., D. Dalevi, F. Levander, and T. Rognvaldsson. 2004. Modular, scriptable and automated analysis tools for high-throughput peptide mass finger-printing. Bioinformatics 20:3628–3635.Crossref, Medline, CAS, Google Scholar8. Patterson, S.D. and R.H. Aebersold. 2003. Proteomics: the first decade and beyond. Nat. Genet. 33:311–323.Crossref, Medline, CAS, Google Scholar9. Anderson, N.L., M. Polanski, R. Pieper, T. Gatlin, R.S. Tirumalai, T.P. Conrads, T.D. Veenstra, J.N. Adkins, et al.. 2004. The human plasma proteome: a nonredundant list developed by combination of four separate sources. Mol. Cell Proteomics 3:311–326.Crossref, Medline, CAS, Google ScholarFiguresReferencesRelatedDetailsCited ByAffinity separation and enrichment methods in proteomic analysisJournal of Proteomics, Vol. 71, No. 3 Vol. 38, No. 6S Follow us on social media for the latest updates Metrics Downloaded 216 times History Published online 30 May 2018 Published in print June 2005 Information© 2005 Author(s)AcknowledgmentsAffymex is available at the TOOLS section on the web site main menu of GenWay Biotech (www.genwaybio.com).Competing Interests StatementThe authors declare that they are seeking a patent on the method and related applications described.PDF download