An insight into high-resolution mass-spectrometry data

Controlling for false discoveries subsequently to large scale one-way ANOVA testing in proteomics: Practical considerations

In discovery proteomics, as well as many other "omic" approaches, the possibility to test for the differential abundance of hundreds (or of thousands) of features simultaneously is appealing, despite requiring specific statistical safeguards, among which controlling for the false discovery rate (FDR) has become standard. Moreover, when more than two biological conditions or group treatments are considered, it has become customary to rely on the one-way analysis of variance (ANOVA) framework, where a first global differential abundance landscape provided by an omnibus test can be subsequently refined using various post-hoc tests (PHTs). However, the interactions between the FDR control procedures and the PHTs are complex, because both correspond to different types of multiple test corrections (MTCs). This article surveys various ways to orchestrate them in a data processing workflow and discusses their pros and cons.

The plasma peptides of sepsis

Background

A practical strategy to discover sepsis specific proteins may be to compare the plasma peptides and proteins from patients in the intensive care unit with and without sepsis. The aim was to discover proteins and/or peptides that show greater observation frequency and/or precursor intensity in sepsis. The endogenous tryptic peptides of ICU-Sepsis were compared to ICU Control, ovarian cancer, breast cancer, female normal, sepsis, heart attack, Alzheimer’s and multiple sclerosis along with their institution-matched controls, female normals and normal samples collected directly onto ice.

Methods

Endogenous tryptic peptides were extracted from individual sepsis and control EDTA plasma samples in a step gradient of acetonitrile for random and independent sampling by LC–ESI–MS/MS with a set of robust and sensitive linear quadrupole ion traps. The MS/MS spectra were fit to fully tryptic peptides within proteins using the X!TANDEM algorithm. The protein observation frequency was counted using the SEQUEST algorithm after selecting the single best charge state and peptide sequence for each MS/MS spectra. The protein observation frequency of ICU-sepsis versus ICU Control was subsequently tested by Chi square analysis. The average protein or peptide log₁₀ precursor intensity was compared across disease and control treatments by ANOVA in the R statistical system.

Results

Peptides and/or phosphopeptides of common plasma proteins such as ITIH3, SAA2, SAA1, and FN1 showed increased observation frequency by Chi square (χ² > 9, p < 0.003) and/or precursor intensity in sepsis. Cellular gene symbols with large Chi square values from tryptic peptides included POTEB, CTNNA1, U2SURP, KIF24, NLGN2, KSR1, GTF2H1, KIT, RPS6KL1, VAV2, HSPA7, SMC2, TCEB3B, ZNF300, SUPV3L1, ADAMTS20, LAMB4, MCCC1, SUPT6H, SCN9A, SBNO1, EPHA1, ABLIM2, cB5E3.2, EPHA10, GRIN2B, HIVEP2, CCL16, TKT, LRP2 and TMF1 amongst others showed increased observation frequency. Similarly, increased frequency of tryptic phosphopeptides were observed from POM121C, SCN8A, TMED8, NSUN7, SLX4, MADD, DNLZ, PDE3B, UTY, DEPDC7, MTX1, MYO1E, RXRB, SYDE1, FN1, PUS7L, FYCO1, USP26, ACAP2, AHI1, KSR2, LMAN1, ZNF280D and SLC8A2 amongst others. Increases in mean precursor intensity in peptides from common plasma proteins such as ITIH3, SAA2, SAA1, and FN1 as well as cellular proteins such as COL24A1, POTEB, KANK1, SDCBP2, DNAH11, ADAMTS7, MLLT1, TTC21A, TSHR, SLX4, MTCH1, and PUS7L among others were associated with sepsis. The processing of SAA1 included the cleavage of the terminal peptide D/PNHFRPAGLPEKY from the most hydrophilic point of SAA1 on the COOH side of the cystatin C binding that was most apparent in ICU-Sepsis patients compared to all other diseases and controls. Additional cleavage of SAA1 on the NH2 terminus side of the cystatin binding site were observed in ICU-Sepsis. Thus there was disease associated variation in the processing of SAA1 in ICU-Sepsis versus ICU controls or other diseases and controls.

Conclusion

Specific proteins and peptides that vary between diseases might be discovered by the random and independent sampling of multiple disease and control plasma from different hospital and clinics by LC–ESI–MS/MS for storage in a relational SQL Server database and analysis with the R statistical system that will be a powerful tool for clinical research. The processing of SAA1 may play an unappreciated role in the inflammatory response to Sepsis.

The plasma peptides of breast versus ovarian cancer

Background

There is a need to demonstrate a proof of principle that proteomics has the capacity to analyze plasma from breast cancer versus other diseases and controls in a multisite clinical trial design. The peptides or proteins that show a high observation frequency, and/or precursor intensity, specific to breast cancer plasma might be discovered by comparison to other diseases and matched controls. The endogenous tryptic peptides of breast cancer plasma were compared to ovarian cancer, female normal, sepsis, heart attack, Alzheimer's and multiple sclerosis along with the institution-matched normal and control samples collected directly onto ice.

Methods

Endogenous tryptic peptides were extracted from individual breast cancer and control EDTA plasma samples in a step gradient of acetonitrile, and collected over preparative C18 for LC-ESI-MS/MS with a set of LTQ XL linear quadrupole ion traps working together in parallel to randomly and independently sample clinical populations. The MS/MS spectra were fit to fully tryptic peptides or phosphopeptides within proteins using the X!TANDEM algorithm. The protein observation frequency was counted using the SEQUEST algorithm after selecting the single best charge state and peptide sequence for each MS/MS spectra. The observation frequency was subsequently tested by Chi Square analysis. The log₁₀ precursor intensity was compared by ANOVA in the R statistical system.

Results

Peptides and/or phosphopeptides of common plasma proteins such as APOE, C4A, C4B, C3, APOA1, APOC2, APOC4, ITIH3 and ITIH4 showed increased observation frequency and/or precursor intensity in breast cancer. Many cellular proteins also showed large changes in frequency by Chi Square (χ² > 100, p < 0.0001) in the breast cancer samples such as CPEB1, LTBP4, HIF-1A, IGHE, RAB44, NEFM, C19orf82, SLC35B1, 1D12A, C8orf34, HIF1A, OCLN, EYA1, HLA-DRB1, LARS, PTPDC1, WWC1, ZNF562, PTMA, MGAT1, NDUFA1, NOGOC, OR1E1, OR1E2, CFI, HSA12, GCSH, ELTD1, TBX15, NR2C2, FLJ00045, PDLIM1, GALNT9, ASH2L, PPFIBP1, LRRC4B, SLCO3A1, BHMT2, CS, FAM188B2, LGALS7, SAT2, SFRS8, SLC22A12, WNT9B, SLC2A4, ZNF101, WT1, CCDC47, ERLIN1, SPFH1, EID2, THOC1, DDX47, MREG, PTPRE, EMILIN1, DKFZp779G1236 and MAP3K8 among others. The protein gene symbols with large Chi Square values were significantly enriched in proteins that showed a complex set of previously established functional and structural relationships by STRING analysis. An increase in mean precursor intensity of peptides was observed for QSER1 as well as SLC35B1, IQCJ-SCHIP1, MREG, BHMT2, LGALS7, THOC1, ANXA4, DHDDS, SAT2, PTMA and FYCO1 among others. In contrast, the QSER1 peptide QPKVKAEPPPK was apparently specific to ovarian cancer.

Conclusion

There was striking agreement between the breast cancer plasma peptides and proteins discovered by LC-ESI-MS/MS with previous biomarkers from tumors, cells lines or body fluids by genetic or biochemical methods. The results indicate that variation in plasma peptides from breast cancer versus ovarian cancer may be directly discovered by LC-ESI-MS/MS that will be a powerful tool for clinical research. It may be possible to use a battery of sensitive and robust linear quadrupole ion traps for random and independent sampling of plasma from a multisite clinical trial.

Millimeter/Submillimeter Spectroscopic Detection of Desorbed Ices: A New Technique in Laboratory Astrochemistry

A new laboratory technique has been developed that utilizes gas-phase, direct-absorption millimeter and submillimeter spectroscopy to detect and identify desorbed species from interstellar and cometary ice analogues. Rotational spectroscopy is a powerful structure-specific technique for detecting isomers and other species possessing the same mass that are indistinguishable with mass spectrometry. Furthermore, the resultant laboratory spectra are directly comparable to observational data from far-infrared and millimeter telescopes. Here, we present the proof-of-concept measurements of the detection of thermally desorbed H₂O, D₂O, and CH₃OH originating in a solid film created at low temperature (∼12 K). The surface binding energy of H₂O is reported and compared to results from traditional techniques, including mass spectrometry and quartz-crystal microbalance measurements of mass loss. Lastly, we demonstrate that this technique can be used to derive thermodynamic values including the sublimation enthalpy and entropy of H₂O.

3D biomaterial matrix to support long term, full thickness, immuno-competent human skin equivalents with nervous system components

Current commercially available human skin equivalents (HSEs) are used for relatively short term studies (∼1 week) due in part to the time-dependent contraction of the collagen gel-based matrix and the limited cell types and skin tissue components utilized. In contrast, here we describe a new matrix consisting of a silk-collagen composite system that provides long term, stable cultivation with reduced contraction and degradation over time. This matrix supports full thickness skin equivalents which include nerves. The unique silk-collagen composite system preserves cell-binding domains of collagen while maintaining the stability and mechanics of the skin system for long-term culture with silk. The utility of this new composite protein-based biomaterial was demonstrated by bioengineering full thickness human skin systems using primary cells, including nerves and immune cells to establish an HSE with a neuro-immuno-cutaneous system. The HSEs with neurons and hypodermis, compared to in vitro skin-only HSEs controls, demonstrated higher secretion of pro-inflammatory cytokines. Proteomics analysis confirmed the presence of several proteins associated with inflammation across all sample groups, but HSEs with neurons had the highest amount of detected protein due to the complexity of the model. This improved, in vitro full thickness HSE model system utilizes cross-linked silk-collagen as the biomaterial and allows reduced reliance on animal models and provides a new in vitro tissue system for the assessment of chronic responses related to skin diseases and drug discovery.

The plasma peptides of ovarian cancer

Background

It may be possible to discover new diagnostic or therapeutic peptides or proteins from blood plasma by using liquid chromatography and tandem mass spectrometry to identify, quantify and compare the peptides cleaved ex vivo from different clinical populations. The endogenous tryptic peptides of ovarian cancer plasma were compared to breast cancer and female cancer normal controls, other diseases with their matched or normal controls, plus ice cold plasma to control for pre-analytical variation.

Methods

The endogenous tryptic peptides or tryptic phospho peptides (i.e. without exogenous digestion) were analyzed from 200 μl of EDTA plasma. The plasma peptides were extracted by a step gradient of organic/water with differential centrifugation, dried, and collected over C18 for analytical HPLC nano electrospray ionization and tandem mass spectrometry (LC–ESI–MS/MS) with a linear quadrupole ion trap. The endogenous peptides of ovarian cancer were compared to multiple disease and normal samples from different institutions alongside ice cold controls. Peptides were randomly and independently sampled by LC–ESI–MS/MS. Precursor ions from peptides > E4 counts were identified by the SEQUEST and X!TANDEM algorithms, filtered in SQL Server, before testing of frequency counts by Chi Square (χ²), for analysis with the STRING algorithm, and comparison of precursor intensity by ANOVA in the R statistical system with the Tukey-Kramer Honestly Significant Difference (HSD) test.

Results

Peptides and/or phosphopeptides of common plasma proteins such as HPR, HP, HPX, and SERPINA1 showed increased observation frequency and/or precursor intensity in ovarian cancer. Many cellular proteins showed large changes in frequency by Chi Square (χ² > 60, p < 0.0001) in the ovarian cancer samples such as ZNF91, ZNF254, F13A1, LOC102723511, ZNF253, QSER1, P4HA1, GPC6, LMNB2, PYGB, NBR1, CCNI2, LOC101930455, TRPM5, IGSF1, ITGB1, CHD6, SIRT1, NEFM, SKOR2, SUPT20HL1, PLCE1, CCDC148, CPSF3, MORN3, NMI, XTP11, LOC101927572, SMC5, SEMA6B, LOXL3, SEZ6L2, and DHCR24. The protein gene symbols with large Chi Square values were significantly enriched in proteins that showed a complex set of previously established functional and structural relationships by STRING analysis. Analysis of the frequently observed proteins by ANOVA confirmed increases in mean precursor intensity in ZFN91, TRPM5, SIRT1, CHD6, RIMS1, LOC101930455 (XP_005275896), CCDC37 and GIMAP4 between ovarian cancer versus normal female and other diseases or controls by the Tukey–Kramer HSD test.

Conclusion

Here we show that separation of endogenous peptides with a step gradient of organic/water and differential centrifugation followed by random and independent sampling by LC–ESI–MS/MS with analysis of peptide frequency and intensity by SQL Server and R revealed significant difference in the ex vivo cleavage of peptides between ovarian cancer and other clinical treatments. There was striking agreement between the proteins discovered from cancer plasma versus previous biomarkers discovered in tumors by genetic or biochemical methods. The results indicate that variation in plasma proteins from ovarian cancer may be directly discovered by LC–ESI–MS/MS that will be a powerful tool for clinical research.

Electronic supplementary material

The online version of this article (10.1186/s12014-018-9215-z) contains supplementary material, which is available to authorized users.

The plasma peptidome

Background

It may be possible to discover new diagnostic or therapeutic peptides or proteins from blood plasma using LC–ESI–MS/MS to identify, with a linear quadrupole ion trap to identify, quantify and compare the statistical distributions of peptides cleaved ex vivo from plasma samples from different clinical populations.

Methods

A systematic method for the organic fractionation of plasma peptides was applied to identify and quantify the endogenous tryptic peptides from human plasma from multiple institutions by C18 HPLC followed nano electrospray ionization and tandem mass spectrometry (LC–ESI–MS/MS) with a linear quadrupole ion trap. The endogenous tryptic peptides, or tryptic phospho peptides (i.e. without exogenous digestion), were extracted in a mixture of organic solvent and water, dried and collected by preparative C18. The tryptic peptides from 6 institutions with 12 different disease and normal EDTA plasma populations, alongside ice cold controls for pre-analytical variation, were characterized by mass spectrometry. Each patient plasma was precipitated in 90% acetonitrile and the endogenous tryptic peptides extracted by a stepwise gradient of increasing water and then formic acid resulting in 10 sub-fractions. The fractionated peptides were manually collected over preparative C18 and injected for 1508 LC–ESI–MS/MS experiments analyzed in SQL Server R.

Results

Peptides that were cleaved in human plasma by a tryptic activity ex vivo provided convenient and sensitive access to most human proteins in plasma that show differences in the frequency or intensity of proteins observed across populations that may have clinical significance. Combination of step wise organic extraction of 200 μL of plasma with nano electrospray resulted in the confident identification and quantification ~ 14,000 gene symbols by X!TANDEM that is the largest number of blood proteins identified to date and shows that you can monitor the ex vivo proteolysis of most human proteins, including interleukins, from blood. A total of 15,968,550 MS/MS spectra ≥ E4 intensity counts were correlated by the SEQUEST and X!TANDEM algorithms to a federated library of 157,478 protein sequences that were filtered for best charge state (2+ or 3+) and peptide sequence in SQL Server resulting in 1,916,672 distinct best-fit peptide correlations for analysis with the R statistical system. SEQUEST identified some 140,054 protein accessions, or some ~ 26,000 gene symbols, proteins or loci, with at least 5 independent correlations. The X!TANDEM algorithm made at least 5 best fit correlations to more than 14,000 protein gene symbols with p-values and FDR corrected q-values of ~ 0.001 or less. Log₁₀ peptide intensity values showed a Gaussian distribution from E8 to E4 arbitrary counts by quantile plot, and significant variation in average precursor intensity across the disease and controls treatments by ANOVA with means compared by the Tukey–Kramer test. STRING analysis of the top 2000 gene symbols showed a tight association of cellular proteins that were apparently present in the plasma as protein complexes with related cellular components, molecular functions and biological processes.

Conclusions

The random and independent sampling of pre-fractionated blood peptides by LC-ESI-MS/MS with SQL Server-R analysis revealed the largest plasma proteome to date and was a practical method to quantify and compare the frequency or log₁₀ intensity of individual proteins cleaved ex vivo across populations of plasma samples from multiple clinical locations to discover treatment-specific variation using classical statistics suitable for clinical science. It was possible to identify and quantify nearly all human proteins from EDTA plasma and compare the results of thousands of LC–ESI–MS/MS experiments from multiple clinical populations using standard database methods in SQL Server and classical statistical strategies in the R data analysis system.

Electronic supplementary material

The online version of this article (10.1186/s12014-018-9211-3) contains supplementary material, which is available to authorized users.

Re-evaluation of the rabbit myosin protein standard used to create the empirical statistical model for decoy library searching

A Rabbit myosin standard, like that used to create the empirical statistical model, was randomly and independently sampled by liquid chromatography micro electrospray ionization and tandem mass spectrometry (LC-ESI-MS/MS) with a linear quadrupole ion trap. The rabbit myosin protein standard appeared pure by SDS-PAGE and CBBR staining but showed many other proteins by silver staining. The LC-MS intensity from myosin and IgG samples were above the 99% safe limit of detection and quantification computed from 36 blank LC-ESI-MS/MS runs. The myosin contained ≤406 Gene Symbols, open reading frames or loci where 79 protein types showed ≥3 peptides from X!TANDEM. Myosins, actin, troponin, other proteins showed 95%-100% homology between the rabbit versus the human decoy library. The myosin protein complex from STRING was true positive compared to random or noise spectra MS/MS with a low type I error (p-value) and low FDR (q-value) computed in R. SDS-PAGE, Western blot, comparison to random and noise MS/MS spectra, X!TANDEM p-values, FDR corrected q-values, and STRING all agreed that the error rate of LC-ESI-MS/MS with a quadrupole ion trap is far below that assumed a priori by the design of the empirical statistical model for decoy library searching.

Variance component analysis to assess protein quantification in biomarker validation: application to selected reaction monitoring-mass spectrometry

BACKGROUND

In the field of biomarker validation with mass spectrometry, controlling the technical variability is a critical issue. In selected reaction monitoring (SRM) measurements, this issue provides the opportunity of using variance component analysis to distinguish various sources of variability. However, in case of unbalanced data (unequal number of observations in all factor combinations), the classical methods cannot correctly estimate the various sources of variability, particularly in presence of interaction. The present paper proposes an extension of the variance component analysis to estimate the various components of the variance, including an interaction component in case of unbalanced data.

RESULTS

We applied an experimental design that uses a serial dilution to generate known relative protein concentrations and estimated these concentrations by two processing algorithms, a classical and a more recent one. The extended method allowed estimating the variances explained by the dilution and the technical process by each algorithm in an experiment with 9 proteins: L-FABP, 14.3.3 sigma, Calgi, Def.A6, Villin, Calmo, I-FABP, Peroxi-5, and S100A14. Whereas, the recent algorithm gave a higher dilution variance and a lower technical variance than the classical one in two proteins with three peptides (L-FABP and Villin), there were no significant difference between the two algorithms on all proteins.

CONCLUSIONS

The extension of the variance component analysis was able to estimate correctly the variance components of protein concentration measurement in case of unbalanced design.

A method for the extraction of the endogenous tryptic peptides (peptidome) from human EDTA plasma

The proteins identified from endogenous peptides agreed between serum versus plasma, and tryptic versus non-tryptic peptides, when collected by C18 alone and analyzed by liquid chromatography electrospray ionization and tandem mass spectrometry (LC-ESI-MS/MS) including amyloids, apolipoproteins, haptoglobin, complements, fibrinogens, hemopexin, antitrypsin and alpha 2 macroglobulin. Precipitation of polypeptides from plasma in 9 vol of 100% organic solvent followed by stepwise extraction of the insoluble pellet with an increasing fraction of water identified thousands of proteins. A Coomassie-blue protein binding assay, and tricine SDS-PAGE, showed that Acetonitrile-Water (AH) resulted in a greater relative enrichment of low molecular weight plasma polypeptides than Acetonitrile-Methanol Water (AMH). A total of 905,386 MS/MS spectra greater than ~10,000 (E4) counts were correlated by X!TANDEM to a federated human protein library of 153,124 different protein sequences that resulted in 58,223 fully tryptic peptides from 3463 Gene Symbols of which 1880 had ≥ 5 independent peptides (p ≤ 0.00001). The results were filtered and organized in an SQL database for analysis using the generic R statistical analysis system. Cellular proteins including secreted and exosome proteins, signaling factors, nucleic acid binding proteins, metabolic enzymes and uncharacterized factors were observed with a significant enrichment of expected protein-protein interactions by STRING analysis.

Freeze-dried plasma proteins are stable at room temperature for at least 1 year

Thirty human EDTA plasma samples from male and female subjects ranging in age from 24 to 74 years were collected on ice, processed ice cold and stored frozen at -80 °C, in liquid nitrogen (LN2), or freeze dried and stored at room temperature in a desiccator (FDRT) or freeze dried and stored at -20 °C for 1 year (FD-20). In a separate experiment, EDTA plasma samples were collected onto ice, processed ice cold and maintained on ice ± protease inhibitors versus incubated at room temperature for up to 96 h. Random and independent sampling by liquid chromatography and tandem mass spectrometry (LC-ESI-MS/MS), as correlated by the MASCOT, OMSSA, X!TANDEM and SEQUEST algorithms, showed that tryptic peptides from complement component 4B (C4B) were rapidly released in plasma at room temperature. Random sampling by LC-ESI-MS/MS showed that peptides from C4B were undetectable on ice, but peptides were cleaved from the mature C4B protein including NGFKSHALQLNNR within as little as 1 h at room temperature. The frequency and intensity of precursors within ± 3 m/z of the C4B peptide NGFKSHALQLNNR was confirmed by automated targeted analysis where the precursors from MS/MS spectra that correlated to the target sequence were analyzed in SQL/R. The C4B preproprotein was processed at the N terminus to release the mature chain that was cleaved on the carboxyl side of the isoprene C2 domain within a polar C terminal sequence of the mature C4B protein, to reveal the thioester reaction site, consistent with LC-ESI-MS/MS and Western blot. Random sampling showed that proteolytic peptides from complement component C4B were rarely observed with long term storage at - 80 °C in a freezer or in liquid nitrogen (LN2), freeze drying with storage at - 20 °C (FD-20 °C) or freeze drying and storage at room temperature (FDRT). Plasma samples maintained at room temperature (RT) showed at least 10-fold to 100-fold greater frequency of peptide correlation to C4B and measured peptide intensity compared to samples on ice for up to 72 h or stored at - 80 °C, LN2, FDRT or FD-20 °C for up to a year.

Binary Markov Random Fields and interpretable mass spectra discrimination

For mass spectra acquired from cancer patients by MALDI or SELDI techniques, automated discrimination between cancer types or stages has often been implemented by machine learning algorithms. Nevertheless, these techniques typically lack interpretability in terms of biomarkers. In this paper, we propose a new mass spectra discrimination algorithm by parameterized Markov Random Fields to automatically generate interpretable classifiers with small groups of scored biomarkers. A dataset of 238 MALDI colorectal mass spectra and two datasets of 216 and 253 SELDI ovarian mass spectra respectively were used to test our approach. The results show that our approach reaches accuracies of 81% to 100% to discriminate between patients from different colorectal and ovarian cancer stages, and performs as well or better than previous studies on similar datasets. Moreover, our approach enables efficient planar-displays to visualize mass spectra discrimination and has good asymptotic performance for large datasets. Thus, our classifiers should facilitate the choice and planning of further experiments for biological interpretation of cancer discriminating signatures. In our experiments, the number of mass spectra for each colorectal cancer stage is roughly half of that for each ovarian cancer stage, so that we reach lower discrimination accuracy for colorectal cancer than for ovarian cancer.

Improving protein identification from tandem mass spectrometry data by one-step methods and integrating data from other platforms

MOTIVATION

Many approaches have been proposed for the protein identification problem based on tandem mass spectrometry (MS/MS) data. In these experiments, proteins are digested into peptides and the resulting peptide mixture is subjected to mass spectrometry. Some interesting putative peptide features (peaks) are selected from the mass spectra. Following that, the precursor ions undergo fragmentation and are analyzed by MS/MS. The process of identification of peptides from the mass spectra and the constituent proteins in the sample is called protein identification from MS/MS data. There are many two-step protein identification procedures, reviewed in the literature, which first attempt to identify the peptides in a separate process and then use these results to infer the proteins. However, in recent years, there have been attempts to provide a one-step solution to protein identification, which simultaneously identifies the proteins and the peptides in the sample.

RESULTS

In this review, we briefly introduce the most popular two-step protein identification procedure, PeptideProphet coupled with ProteinProphet. Following that, we describe the difficulties with two-step procedures and review some recently introduced one-step protein/peptide identification procedures that do not suffer from these issues. The focus of this review is on one-step procedures that are based on statistical likelihood-based models, but some discussion of other one-step procedures is also included. We report comparative performances of one-step and two-step methods, which support the overall superiorities of one-step procedures. We also cover some recent efforts to improve protein identification by incorporating other molecular data along with MS/MS data.

Statistical issues in the design and planning of proteomic profiling experiments

The statistical design of a clinical proteomics experiment is a critical part of well-undertaken investigation. Standard concepts from experimental design such as randomization, replication and blocking should be applied in all experiments, and this is possible when the experimental conditions are well understood by the investigator. The large number of proteins simultaneously considered in proteomic discovery experiments means that determining the number of required replicates to perform a powerful experiment is more complicated than in simple experiments. However, by using information about the nature of an experiment and making simple assumptions this is achievable for a variety of experiments useful for biomarker discovery and initial validation.

Statistical methods for quantitative mass spectrometry proteomic experiments with labeling

Mass Spectrometry utilizing labeling allows multiple specimens to be subjected to mass spectrometry simultaneously. As a result, between-experiment variability is reduced. Here we describe use of fundamental concepts of statistical experimental design in the labeling framework in order to minimize variability and avoid biases. We demonstrate how to export data in the format that is most efficient for statistical analysis. We demonstrate how to assess the need for normalization, perform normalization, and check whether it worked. We describe how to build a model explaining the observed values and test for differential protein abundance along with descriptive statistics and measures of reliability of the findings. Concepts are illustrated through the use of three case studies utilizing the iTRAQ 4-plex labeling protocol.

A regularized Hotelling's T2 test for pathway analysis in proteomic studies

Recent proteomic studies have identified proteins related to specific phenotypes. In addition to marginal association analysis for individual proteins, analyzing pathways (functionally related sets of proteins) may yield additional valuable insights. Identifying pathways that differ between phenotypes can be conceptualized as a multivariate hypothesis testing problem: whether the mean vector μ of a p-dimensional random vector X is μ₀ . Proteins within the same biological pathway may correlate with one another in a complicated way, and type I error rates can be inflated if such correlations are incorrectly assumed to be absent. The inflation tends to be more pronounced when the sample size is very small or there is a large amount of missingness in the data, as is frequently the case in proteomic discovery studies. To tackle these challenges, we propose a regularized Hotelling's T² (RHT) statistic together with a non-parametric testing procedure, which effectively controls the type I error rate and maintains good power in the presence of complex correlation structures and missing data patterns. We investigate asymptotic properties of the RHT statistic under pertinent assumptions and compare the test performance with four existing methods through simulation examples. We apply the RHT test to a hormone therapy proteomics data set, and identify several interesting biological pathways for which blood serum concentrations changed following hormone therapy initiation.

Identification and quantification of peptides and proteins secreted from prostate epithelial cells by unbiased liquid chromatography tandem mass spectrometry using goodness of fit and analysis of variance

The proteins secreted by prostate cancer cells (PC3(AR)6) were separated by strong anion exchange chromatography, digested with trypsin and analyzed by unbiased liquid chromatography tandem mass spectrometry with an ion trap. The spectra were matched to peptides within proteins using a goodness of fit algorithm that showed a low false positive rate. The parent ions for MS/MS were randomly and independently sampled from a log-normal population and therefore could be analyzed by ANOVA. Normal distribution analysis confirmed that the parent and fragment ion intensity distributions were sampled over 99.9% of their range that was above the background noise. Arranging the ion intensity data with the identified peptide and protein sequences in structured query language (SQL) permitted the quantification of ion intensity across treatments, proteins and peptides. The intensity of 101,905 fragment ions from 1421 peptide precursors of 583 peptides from 233 proteins separated over 11 sample treatments were computed together in one ANOVA model using the statistical analysis system (SAS) prior to Tukey-Kramer honestly significant difference (HSD) testing. Thus complex mixtures of proteins were identified and quantified with a high degree of confidence using an ion trap without isotopic labels, multivariate analysis or comparing chromatographic retention times.

Variance component analysis of a multi-site study for the reproducibility of multiple reaction monitoring measurements of peptides in human plasma

Background

In the Addona et al. paper (Nature Biotechnology 2009), a large-scale multi-site study was performed to quantify Multiple Reaction Monitoring (MRM) measurements of proteins spiked in human plasma. The unlabeled signature peptides derived from the seven target proteins were measured at nine different concentration levels, and their isotopic counterparts were served as the internal standards.

Methodology/Principal Findings

In this paper, the sources of variation are analyzed by decomposing the variance into parts attributable to specific experimental factors: technical replicates, sites, peptides, transitions within each peptide, and higher-order interaction terms based on carefully built mixed effects models. The factors of peptides and transitions are shown to be major contributors to the variance of the measurements considering heavy (isotopic) peptides alone. For the light (¹²C) peptides alone, in addition to these factors, the factor of study*peptide also contributes significantly to the variance of the measurements. Heterogeneous peptide component models as well as influence analysis identify the outlier peptides in the study, which are then excluded from the analysis. Using a log-log scale transformation and subtracting the heavy/isotopic peptide [internal standard] measurement from the peptide measurements (i.e., taking the logarithm of the peak area ratio in the original scale establishes that), the MRM measurements are overall consistent across laboratories following the same standard operating procedures, and the variance components related to sites, transitions and higher-order interaction terms involving sites have greatly reduced impact. Thus the heavy peptides have been effective in reducing apparent inter-site variability. In addition, the estimates of intercepts and slopes of the calibration curves are calculated for the sub-studies.

Conclusions/Significance

The MRM measurements are overall consistent across laboratories following the same standard operating procedures, and heavy peptides can be used as an effective internal standard for reducing apparent inter-site variability. Mixed effects modeling is a valuable tool in mass spectrometry-based proteomics research.

Mass spectrometry-based proteomics and its potential use in haematological research

In the last decade proteomics has made great progress reaching throughput and comprehensiveness comparable to genomics technologies. Mass spectrometry plays a key role in proteomics and has become an indispensable method for molecular and cellular biology because many cellular changes in response to internal or external stimuli can only be detected at the proteome level. Furthermore, different from genomics which depends on the availability of DNA or RNA, proteomics is not restricted to cellular samples, but also allows the analysis of biological fluids like serum, plasma or urine. This article provides an overview of the recent developments in proteomics techniques useful for haematological research.