Increased proteome coverage by combining PAGE and peptide isoelectric focusing: Comparative study of gel‐based separation approaches

The in‐depth analysis of complex proteome samples requires fractionation of the sample into subsamples prior to LC‐MS/MS in shotgun proteomics experiments. We have established a 3D workflow for shotgun proteomics that relies on protein separation by 1D PAGE, gel fractionation, trypsin digestion, and peptide separation by in‐gel IEF, prior to RP‐HPLC‐MS/MS. Our results show that applying peptide IEF can significantly increase the number of proteins identified from PAGE subfractionation. This method delivers deeper proteome coverage and provides a large degree of flexibility in experimentally approaching highly complex mixtures by still relying on protein separation according to molecular weight in the first dimension.     

Comparative analysis of OFFGel,strong cation exchange with pH gradient,and RP at high pH for first‐dimensional separation of peptides from a membrane‐enriched protein fraction

The analysis and quantitation of membrane proteins have proved challenging for proteomics. Although several approaches have been introduced to complement gel‐based analysis of intact proteins, the literature is rather limited in comparing major emerging approaches. Peptide fractionation using IEF (OFFGel), strong cation exchange HPLC using a pH gradient (SCX‐pG), and RP HPLC at high pH, have been shown to increase peptide and protein identification over classic MudPIT approaches. This article compares these three approaches for first‐dimensional separation of peptides using a detergent phase (Triton X‐114) enriched membrane fraction from mouse cortical brain tissue. Results indicate that RP at high pH (pH 10) was superior for the identification of more peptides and proteins in comparison to the OFFGel or the SCX‐pG approaches. In addition, gene ontology analysis (GOMiner) revealed that RP at high pH (pH 10) successfully identified an increased number of proteins with “membrane” ontology, further confirming its suitability for membrane protein analysis, in comparison to SCX‐pG and OFFGel techniques.     

Multidimensional separation prior to mass spectrometry: Getting closer to the bottom of the iceberg

While prefractionation has previously been shown to improve results in MS analysis, a novel combination provides an additional dimension of separation: protein fractionation by SDS‐PAGE followed by IEF of tryptic peptides before separation by RP‐LC [Atanassov and Urlaub, Proteomics 2013, 13, 2947–2955]. This three‐step separation procedure prior to MS/MS substantially increases proteome coverage and represents a further step toward a more comprehensive analysis of complex proteomes.     

Shotgun proteome analysis utilising mixed mode (reversed phase‐anion exchange chromatography) in conjunction with reversed phase liquid chromatography mass spectrometry analysis

The 2‐D peptide separations employing mixed mode reversed phase anion exchange (MM (RP‐AX)) HPLC in the first dimension in conjunction with RP chromatography in the second dimension were developed and utilised for shotgun proteome analysis. Compared with strong cation exchange (SCX) typically employed for shotgun proteomic analysis, peptide separations using MM (RP‐AX) revealed improved separation efficiency and increased peptide distribution across the elution gradient. In addition, improved sample handling, with no significant reduction in the orthogonality of the peptide separations was observed. The shotgun proteomic analysis of a mammalian nuclear cell lysate revealed additional proteome coverage (2818 versus 1125 unique peptides and 602 versus 238 proteins) using the MM (RP‐AX) compared with the traditional SCX hyphenated to RP‐LC‐MS/MS. The MM analysis resulted in approximately 90% of the unique peptides identified present in only one fraction, with a heterogeneous peptide distribution across all fractions. No clustering of the predominant peptide charge states was observed during the gradient elution. The application of MM (RP‐AX) for 2‐D LC proteomic studies was also extended in the analysis of iTRAQ‐labelled HeLa and cyanobacterial proteomes using nano‐flow chromatography interfaced to the MS/MS. We demonstrate MM (RP‐AX) HPLC as an alternative approach for shotgun proteomic studies that offers significant advantages over traditional SCX peptide separations.     

Evaluation on the effect of different in-gel peptide isoelectric focusing parameters in global proteomic profiling

Peptide isoelectric focusing (IEF) is a common technique used in two-dimensional liquid chromatography tandem mass spectrometry (2D–LC–MS/MS) proteomic workflow, in which the tryptic peptide is first pre-fractionated based on pI values before being subjected to reverse phase LC–MS analysis. Although this method has been widely used by many research groups, a systemic study on the optimal conditions and fundamental parameters influencing the experimental outcomes has been lacking, including the effect of peptide extraction methods, the extent of pre-fractionation, and the choice of pH range. In this study, we compared the effect of different parameters on the numbers of peptides and proteins identified using two complex mouse proteomes. The results indicated that extraction of peptides from immobilized pH gradient (IPG) strips by sequential elution of increasingly organic solvents provided the highest number of peptide identification. In addition, we showed that approximately 45 more unique proteins were identified for every additional fraction collected during peptide IEF. Although narrow pH ranges provided higher resolution in peptide separation as expected, different pH ranges yielded similar numbers of peptide and protein identification. Overall, we demonstrated that the extraction solvent influenced the numbers of peptide and protein identification and quantitatively demonstrated the advantage of extensive fractionation and the performance of different pH ranges in practice.     

ICPLQuant – A software for non‐isobaric isotopic labeling proteomics

The main goal of many proteomics experiments is an accurate and rapid quantification and identification of regulated proteins in complex biological samples. The bottleneck in quantitative proteomics remains the availability of efficient software to evaluate and quantify the tremendous amount of mass spectral data acquired during a proteomics project. A new software suite, ICPLQuant, has been developed to accurately quantify isotope‐coded protein label (ICPL)‐labeled peptides on the MS level during LC‐MALDI and peptide mass fingerprint experiments. The tool is able to generate a list of differentially regulated peptide precursors for subsequent MS/MS experiments, minimizing time‐consuming acquisition and interpretation of MS/MS data. ICPLQuant is based on two independent units. Unit 1 performs ICPL multiplex detection and quantification and proposes peptides to be identified by MS/MS. Unit 2 combines MASCOT MS/MS protein identification with the quantitative data and produces a protein/peptide list with all the relevant information accessible for further data mining. The accuracy of quantification, selection of peptides for MS/MS‐identification and the automated output of a protein list of regulated proteins are demonstrated by the comparative analysis of four different mixtures of three proteins (Ovalbumin, Horseradish Peroxidase and Rabbit Albumin) spiked into the complex protein background of the DGPF Proteome Marker.     

Identification of peptide‐binding sites within BSA using rapid,laser‐induced covalent cross‐linking combined with high‐performance mass spectrometry

We are developing a rapid, time‐resolved method using laser‐activated cross‐linking to capture protein‐peptide interactions as a means to interrogate the interaction of serum proteins as delivery systems for peptides and other molecules. A model system was established to investigate the interactions between bovine serum albumin (BSA) and 2 peptides, the tridecapeptide budding‐yeast mating pheromone (α‐factor) and the decapeptide human gonadotropin‐releasing hormone (GnRH). Cross‐linking of α‐factor, using a biotinylated, photoactivatable p‐benzoyl‐L‐phenylalanine (Bpa)–modified analog, was energy‐dependent and achieved within seconds of laser irradiation. Protein blotting with an avidin probe was used to detect biotinylated species in the BSA‐peptide complex. The cross‐linked complex was trypsinized and then interrogated with nano‐LC–MS/MS to identify the peptide cross‐links. Cross‐linking was greatly facilitated by Bpa in the peptide, but some cross‐linking occurred at higher laser powers and high concentrations of a non‐Bpa–modified α‐factor. This was supported by experiments using GnRH, a peptide with sequence homology to α‐factor, which was likewise found to be cross‐linked to BSA by laser irradiation. Analysis of peptides in the mass spectra showed that the binding site for both α‐factor and GnRH was in the BSA pocket defined previously as the site for fatty acid binding. This model system validates the use of laser‐activation to facilitate cross‐linking of Bpa‐containing molecules to proteins. The rapid cross‐linking procedure and high performance of MS/MS to identify cross‐links provides a method to interrogate protein‐peptide interactions in a living cell in a time‐resolved manner.     

Use of narrow‐range peptide IEF to improve detection of lung adenocarcinoma markers in plasma and pleural effusion

In this study we applied narrow‐range peptide IEF to plasma or pleural effusion prior to LC/MS/MS. Two methods for narrow‐range IEF were run; IPG strips and free‐flow electrophoresis. Data from this study was compared with cell line data to evaluate the method performance in body fluids. To test the methods potential in quantitative biomarker discovery studies, plasma and pleural effusion from patients with lung adenocarcinoma (n=3) were compared with inflammatory pleuritis (n=3) using iTRAQ quantification. Using narrow‐range IEF on the peptide level we were able to identify and quantify 282 proteins in plasma and 300 proteins in pleural effusion. These body fluid proteomes demonstrated high degree of overlap; however, more proteins significantly differently altered levels related to adenenocarcinoma were found in pleural effusion compared with plasma, suggesting enrichment of lung tissue‐related proteins in pleural effusion. Nine proteins were chosen for initial validation with Western blot, and one protein (NPC2) was chosen for further validation using imunohistochemistry. Overall, the quantitative results from IEF/LC/MS/MS showed good correlation with the results from Western blot and imunohistochemistry, showing the potential of this methodology in quantitative biomarker discovery studies.     

Identification of cross‐reactive B‐cell epitopes between Bos d 9.0101(Bos Taurus) and Gly m 5.0101 (Glycine max) by epitope mapping MALDI‐TOF MS

Exposure to cow's milk constitutes one of the most common causes of food allergy. In addition, exposure to soy proteins has become relevant in a restricted proportion of milk allergic pediatric patients treated with soy formulae as a dairy substitute, because of the cross‐allergenicity described between soy and milk proteins. We have previously identified several cross‐reactive allergens between milk and soy that may explain this intolerance. The purpose of the present work was to identify epitopes in the purified αS1‐casein and the recombinant soy allergen Gly m 5.0101 (Gly m 5) using an α‐casein‐specific monoclonal antibody (1D5 mAb) through two different approaches for epitope mapping, to understand cross‐reactivity between milk and soy. The 1D5 mAb was immobilized onto magnetic beads, incubated with the peptide mixture previously obtained by enzymatic digestion of the allergens, and the captured peptides were identified by MALDI‐TOF MS analysis. On a second approach, the peptide mixture was resolved by RP‐HPLC and immunodominant peptides were identified by dot blot with the mAb. Finally, recognized peptides were sequenced by MALDI‐TOF MS. This novel MS based approach led us to identify and characterize four peptides on α‐casein and three peptides on Gly m 5 with a common core motif. Information obtained from these cross‐reactive epitopes allows us to gain valuable insight into the molecular mechanisms of cross‐reactivity, to further develop new and more effective vaccines for food allergy.     

Evaluation of the accuracy of protein quantification using isotope TMPP‐labeled peptides

N‐succinimidyloxycarbonylmethyl tris(2,4,6‐trimethoxyphenyl) phosphonium bromide (TMPP‐Ac‐OSu) reacts rapidly, mildly, and specifically with the N‐terminals of proteins and peptides. Thus, it can be developed as an ideal isotope‐coded tag to be used in quantitative proteomics. Here, we present a strategy for light and heavy TMPP‐based quantitative proteomic analysis, in which peptides in a mixture can be quantified using an on‐tip TMPP derivatization approach. To demonstrate the accuracy of this strategy, light and heavy TMPP‐labeled peptides were combined at different ratios and subsequently analyzed by LC‐MS/MS. The MS spectra and scatter plots show that peptide and protein ratios were both consistent with the mixed ratios. We observed a linear correlation between protein ratios and the predicted ratios. In comparison with SILAC method, the TMPP labeling method produced similarly accurate quantitative results with low CVs. In conclusion, our results suggest that this isotope‐coded TMPP method achieved accurate quantification and compatibility with IEF‐based separation. With the inherent advantages of TMPP derivatization, we believe that it holds great promise for future applications in quantitative proteomics analysis.     

Improved accuracy of cell surface shaving proteomics in Staphylococcus aureus using a false‐positive control

Proteolytic treatment of intact bacterial cells is an ideal means for identifying surface‐exposed peptide epitopes and has potential for the discovery of novel vaccine targets. Cell stability during such treatment, however, may become compromised and result in the release of intracellular proteins that complicate the final analysis. Staphylococcus aureus is a major human pathogen, causing community and hospital‐acquired infections, and is a serious healthcare concern due to the increasing prevalence of multiple antibiotic resistances amongst clinical isolates. We employed a cell surface “shaving” technique with either trypsin or proteinase‐K combined with LC‐MS/MS. Trypsin‐derived data were controlled using a “false‐positive” strategy where cells were incubated without protease, removed by centrifugation and the resulting supernatants digested. Peptides identified in this fraction most likely result from cell lysis and were removed from the trypsin‐shaved data set. We identified 42 predicted S. aureus COL surface proteins from 260 surface‐exposed peptides. Trypsin and proteinase‐K digests were highly complementary with ten proteins identified by both, 16 specific to proteinase‐K treatment, 13 specific to trypsin and three identified in the control. The use of a subtracted false‐positive strategy improved enrichment of surface‐exposed peptides in the trypsin data set to approximately 80% (124/155 peptides). Predominant surface proteins were those associated with methicillin resistance–surface protein SACOL0050 (pls) and penicillin‐binding protein 2′ (mecA), as well as bifunctional autolysin and the extracellular matrix‐binding protein Ebh. The cell shaving strategy is a rapid method for identifying surface‐exposed peptide epitopes that may be useful in the design of novel vaccines against S. aureus.     

Comparison of fractionation proteomics for local SWATH library building

For data‐independent acquisition by means of sequential window acquisition of all theoretical fragment ion spectra (SWATH), a reference library of data‐dependent acquisition (DDA) runs is typically used to correlate the quantitative data from the fragment ion spectra with peptide identifications. The quality and coverage of such a reference library is therefore essential when processing SWATH data. In general, library sizes can be increased by reducing the impact of DDA precursor selection with replicate runs or fractionation. However, these strategies can affect the match between the library and SWATH measurement, and thus larger library sizes do not necessarily correspond to improved SWATH quantification. Here, three fractionation strategies to increase local library size were compared to standard library building using replicate DDA injection: protein SDS‐PAGE fractionation, peptide high‐pH RP‐HPLC fractionation and MS‐acquisition gas phase fractionation. The impact of these libraries on SWATH performance was evaluated in terms of the number of extracted peptides and proteins, the match quality of the peptides and the extraction reproducibility of the transitions. These analyses were conducted using the hydrophilic proteome of differentiating human embryonic stem cells. Our results show that SWATH quantitative results and interpretations are affected by choice of fractionation technique. Data are available via ProteomeXchange with identifier PXD006190.     

Protein identification and quantification from riverbank grape,Vitis riparia: Comparing SDS‐PAGE and FASP‐GPF techniques for shotgun proteomic analysis

Protein sample preparation optimisation is critical for establishing reproducible high throughput proteomic analysis. In this study, two different fractionation sample preparation techniques (in‐gel digestion and in‐solution digestion) for shotgun proteomics were used to quantitatively compare proteins identified in Vitis riparia leaf samples. The total number of proteins and peptides identified were compared between filter aided sample preparation (FASP) coupled with gas phase fractionation (GPF) and SDS‐PAGE methods. There was a 24% increase in the total number of reproducibly identified proteins when FASP‐GPF was used. FASP‐GPF is more reproducible, less expensive and a better method than SDS‐PAGE for shotgun proteomics of grapevine samples as it significantly increases protein identification across biological replicates. Total peptide and protein information from the two fractionation techniques is available in PRIDE with the identifier PXD001399 ( http://proteomecentral.proteomexchange.org/dataset/PXD001399 ).     

Extending the coverage of spectral libraries: A neighbor‐based approach to predicting intensities of peptide fragmentation spectra

Searching spectral libraries in MS/MS is an important new approach to improving the quality of peptide and protein identification. The idea relies on the observation that ion intensities in an MS/MS spectrum of a given peptide are generally reproducible across experiments, and thus, matching between spectra from an experiment and the spectra of previously identified peptides stored in a spectral library can lead to better peptide identification compared to the traditional database search. However, the use of libraries is greatly limited by their coverage of peptide sequences: even for well‐studied organisms a large fraction of peptides have not been previously identified. To address this issue, we propose to expand spectral libraries by predicting the MS/MS spectra of peptides based on the spectra of peptides with similar sequences. We first demonstrate that the intensity patterns of dominant fragment ions between similar peptides tend to be similar. In accordance with this observation, we develop a neighbor‐based approach that first selects peptides that are likely to have spectra similar to the target peptide and then combines their spectra using a weighted K‐nearest neighbor method to accurately predict fragment ion intensities corresponding to the target peptide. This approach has the potential to predict spectra for every peptide in the proteome. When rigorous quality criteria are applied, we estimate that the method increases the coverage of spectral libraries available from the National Institute of Standards and Technology by 20–60%, although the values vary with peptide length and charge state. We find that the overall best search performance is achieved when spectral libraries are supplemented by the high quality predicted spectra.     

Comparison of protein and peptide prefractionation methods for the shotgun proteomic analysis of Synechocystis sp. PCC 6803

Proteome analysis by gel-free "shotgun" proteomics relies on the simplification of a peptide mixture before it is analyzed in a mass spectrometer. While separation on a reverse-phase (RP) liquid chromatographic column is widely employed, a variety of other methods have been used to fractionate both proteins and peptides before this step. We compared six different protein and peptide fractionation workflows, using Synechocystis sp. PCC 6803, a useful model cyanobacterium for potential exploitation to improve its production of hydrogen and other secondary metabolites. Pre-digestion protein separation was performed by strip-based isoelectric focusing, one-dimensional polyacrylamide gel electrophoresis, or weak anion exchange chromatography, while pre-RP peptide separation was accomplished by isoelectric focusing (IEF) or strong cation exchange chromatography. Peptides were identified using electrospray ionization quadrupole time of flight-tandem mass spectrometry. Mass spectrometry (MS) and tandem mass spectra were analyzed using ProID software employing both a single organism database and the entire NCBI non-redundant database, and a total of 776 proteins were identified using a stringent set of selection criteria. Method comparisons were made on the basis of the results obtained (number and types of proteins identified), as well as ease of use and other practical aspects. IEF-IEF protein and peptide fractionation prior to RP gave the best overall performance.     

Epitope motif of an anti‐nitrotyrosine antibody specific for tyrosine‐nitrated peptides revealed by a combination of affinity approaches and mass spectrometry

Nitration of tyrosine residues has been shown to be an important oxidative modification in proteins and has been suggested to play a role in several diseases such as atherosclerosis, asthma, lung and neurodegenerative diseases. Detection of nitrated proteins has been mainly based on the use of nitrotyrosine‐specific antibodies. In contrast, only a small number of nitration sites in proteins have been unequivocally identified by MS. We have used a monoclonal 3‐NT‐specific antibody, and have synthesized a series of tyrosine‐nitrated peptides of prostacyclin synthase (PCS) in which a single specific nitration site at Tyr‐430 had been previously identified upon reaction with peroxynitrite 17 . The determination of antibody‐binding affinity and specificity of PCS peptides nitrated at different tyrosine residues (Tyr‐430, Tyr‐421, Tyr‐83) and sequence mutations around the nitration sites provided the identification of an epitope motif containing positively charged amino acids (Lys and/or Arg) N‐terminal to the nitration site. The highest affinity to the anti‐3NT‐antibody was found for the PCS peptide comprising the Tyr‐430 nitration site with a K_D of 60 nM determined for the peptide, PCS(424‐436‐Tyr‐430NO₂); in contrast, PCS peptides nitrated at Tyr‐421 and Tyr‐83 had substantially lower affinity. ELISA, SAW bioaffinity, proteolytic digestion of antibody‐bound peptides and affinity‐MS analysis revealed highest affinity to the antibody for tyrosine‐nitrated peptides that contained positively charged amino acids in the N‐terminal sequence to the nitration site. Remarkably, similar N‐terminal sequences of tyrosine‐nitration sites have been recently identified in nitrated physiological proteins, such as eosinophil peroxidase and eosinophil‐cationic protein. Copyright © 2011 European Peptide Society and John Wiley & Sons, Ltd.     

Analysis of the phosphoproteome of the multicellular bacterium Streptomyces coelicolor A3(2) by protein/peptide fractionation,phosphopeptide enrichment and high‐accuracy mass spectrometry

The serine (Ser)/threonine (Thr)/tyrosine (Tyr) phosphoproteome of exponentially growing Streptomyces coelicolor A3(2) was analysed using the gel‐free approaches of preparative IEF for protein fractionation, followed by strong cation exchange peptide fractionation and phosphopeptide enrichment by TiO₂ metal oxide affinity chromatography. Phosphopeptides were identified using LC‐ESI‐LTQ‐Orbitrap^? MS. Forty‐six novel phosphorylation sites were identified on 40 proteins involved in gene regulation or signalling, central metabolism, protein biosynthesis, membrane transport and cell division, as well as several of unknown function. In contrast to other studies, Thr phosphorylation appeared to be preferred, with relative levels of Ser, Thr and Tyr phosphorylation of 34, 52 and 14%, respectively. Genes for most of the 40 phosphorylated proteins reside in the central “housekeeping” region of the linear S. coelicolor chromosome, suggesting that in general Ser, Thr and Tyr phosphorylation play a role in regulating essential aspects of metabolism in streptomycetes. A greater number of regulators and putative regulators were also identified compared with other bacterial phosphoproteome studies, potentially reflecting the complex heterotrophic and developmental life style of S. coelicolor. This study is the first analysis of the phosphoproteome of a member of this morphologically complex and industrially important group of microorganisms.     

DeepQuanTR: MALDI‐MS‐based label‐free quantification of proteins in complex biological samples

The quantification of changes in protein abundance in complex biological specimens is essential for proteomic studies in basic and applied research. Here we report on the development and validation of the DeepQuanTR software for identification and quantification of differentially expressed proteins using LC‐MALDI‐MS. Following enzymatic digestion, HPLC peptide separation and normalization of MALDI‐MS signal intensities to the ones of internal standards, the software extracts peptide features, adjusts differences in HPLC retention times and performs a relative quantification of features. The annotation of multiple peptides to the corresponding parent protein allows the definition of a Protein Quant Value, which is related to protein abundance and which allows inter‐sample comparisons. The performance of DeepQuanTR was evaluated by analyzing 24 samples deriving from human serum spiked with different amounts of four proteins and eight complex samples of vascular proteins, derived from surgically resected human kidneys with cancer following ex vivo perfusion with a reactive ester biotin derivative. The identification and experimental validation of proteins, which were differentially regulated in cancerous lesions as compared with normal kidney, was used to demonstrate the power of DeepQuanTR. This software, which can easily be used with established proteomic methodologies, facilitates the relative quantification of proteins derived from a wide variety of different samples.     

Identification and characterization of the Sulfolobus solfataricus P2 proteome

Via combined separation approaches, a total of 1399 proteins were identified, representing 47% of the Sulfolobus solfataricus P2 theoretical proteome. This includes 1323 proteins from the soluble fraction, 44 from the insoluble fraction and 32 from the extra-cellular or secreted fraction. We used conventional 2-dimensional gel electrophoresis (2-DE) for the soluble fraction, and shotgun proteomics for all three cell fractions (soluble, insoluble, and secreted). Two gel-based fractionation methods were explored for shotgun proteomics, namely: (i) protein separation utilizing 1-dimensional gel electrophoresis (1-DE) followed by peptide fractionation by iso-electric focusing (IEF), and (ii) protein and peptide fractionation both employing IEF. Results indicate that a 1D-IEF fractionation workflow with three replicate mass spectrometric analyses gave the best overall result for soluble protein identification. A greater than 50% increment in protein identification was achieved with three injections using LC-ESI-MS/MS. Protein and peptide fractionation efficiency; together with the filtration criteria are also discussed.     

Gel‐free mass spectrometry analysis of Drosophila melanogaster heads

Membrane proteins play key roles in several fundamental biological processes such as cell signalling, energy metabolism and transport. Despite the significance, these still remain an under‐represented group in proteomics datasets. Herein, a bottom‐up approach to analyse an enriched membrane fraction from Drosophila melanogaster heads using multidimensional liquid chromatography (LC) coupled with tandem‐mass spectrometry (MS/MS) that relies on complete solubilisation and digestion of proteins, is reported. An enriched membrane fraction was prepared using equilibrium density centrifugation on a discontinuous sucrose gradient, followed by solubilisation using the filter‐aided sample preparation (FASP), tryptic and sequential chymotryptic digestion of proteins. Peptides were separated by reversed‐phase (RP) LC at high pH in the first dimension and acidic RP‐LC in the second dimension coupled directly to an Orbitrap Velos Pro mass spectrometer. A total number of 4812 proteins from 114 865 redundant and 38 179 distinct peptides corresponding to 4559 genes were identified in the enriched membrane fraction from fly heads. These included brain receptors, transporters and channels that are most important elements as drug targets or are linked to disease. Data are available via ProteomeXchange with identifier PXD001712 ( http://proteomecentral.proteomexchange.org/dataset/PXD001712 ).