Prediction of peptide retention times in high-pressure liquid chromatography on the basis of amino acid composition

Correctness of protein identifications of Bacillus subtilis proteome with the indication on potential false positive peptides supported by predictions of their retention times

The predictive capability of the retention time prediction model based on quantitative structure-retention relationships (QSRR) was tested. QSRR model was derived with the use of set of peptides identified with the highest scores and originated from 8 known proteins annotated as model ones. The predictive ability of the QSRR model was verified with the use of a Bacillus subtilis proteome digest after separation and identification of the peptides by LC-ESI-MS/MS. That ability was tested with three sets of testing peptides assigned to the proteins identified with different levels of confidence. First, the set of peptides identified with the highest scores achieved in the search were considered. Hence, proteins identified on the basis of more than one peptide were taken into account. Furthermore, proteins identified on the basis of just one peptide were also considered and, depending on the possessed scores, both above and below the assumed threshold, were analyzed in two separated sets. The QSRR approach was applied as the additional constraint in proteomic research verifying results of MS/MS ion search and confirming the correctness of the peptides identifications along with the indication of the potential false positives.

emPAI Calc--for the estimation of protein abundance from large-scale identification data by liquid chromatography-tandem mass spectrometry

SUMMARY

emPAI Calc is an open-source web application for the estimation of protein abundance. It uses the correlation between the number of identified peptides and protein abundance in mass spectrometry-based proteomic experiments. The program is the first implementation of our previously reported emPAI algorithm; it calculates the emPAI from the protein identification results obtained by database search engines such as Mascot.

Unambiguous determination of isobaric histone modifications by reversed-phase retention time and high-mass accuracy

Methylation and acetylation of lysines are crucial posttranslational modifications that regulate gene transcription and have been shown to be misregulated in many forms of cancers. Western blot, immunoprecipitation, and immunofluorescence are commonly used to characterize histone acetylation and methylation. However, these approaches are limited by the availability, site specificity, and cross-reactivity of antibodies. Mass spectrometry is emerging as an additional powerful tool for histone characterization. The isobaric nature of trimethylation and acetylation (42.0470 and 42.0106 Da, respectively) confounds histone characterization by means other than high-resolution/high-mass accuracy mass spectrometry. In this study, we adapted methodology that exploits difference in the relative retention time of acetylated and methylated peptides to unequivocally distinguish between these two modifications even with low-mass accuracy mass spectrometers. The approach was tested on tryptic digest of Saccharomyces cerevisiae histones. We found that acetylation resulted in increased retention in reversed-phase chromatography, whereas methylation, including trimethylation, showed little change in retention. For example, the acetylated forms of peptide (27)KSAPSTGGVKKPHR(40) eluted at 15.63 min, whereas the methylated forms eluted at 13.89 min. In addition, the effect of acetylation was cumulative as observed in the case of peptide (9)KSTGGKAPR(17), whose unmodified, monoacetylated, and diacetylated isoforms eluted at 7.43, 10.47, and 16.49 min, respectively. The modification patterns of the peptides in question were subsequently verified by high-mass accuracy tandem mass spectrometry.

The distribution of CTL epitopes in HIV-1 appears to be random, and similar to that of other proteomes

Background

HIV-1 viruses are highly capable of mutating their proteins to escape the presentation of CTL epitopes in their current host. Upon transmission to another host, some escape mutations revert, but other remain stable in the virus sequence for at least several years. Depending on the rate of accumulation and reversion of escape mutations, HIV-1 could reach a high level of adaptation to the human population. Yusim et. al. hypothesized that the apparent clustering of CTL epitopes in the conserved regions of HIV-1 proteins could be an evolutionary signature left by large-scale adaptation of HIV-1 to its human/simian host.

Results

In this paper we quantified the distribution of CTL epitopes in HIV-1 and found that that in 99% of the HIV-1 protein sequences, the epitope distribution was indistinguishable from random. Similar percentages were found for HCV, Influenza and for three eukaryote proteomes (Human, Drosophila, Yeast).

Conclusion

We conclude that CTL epitopes in HIV-1 are randomly distributed, and that this distribution is similar to the distribution of CTL epitopes in proteins from other proteomes. Therefore, the visually apparent clustering of CTL epitopes in epitope maps should not be interpreted as a signature of a past large-scale adaptation of HIV-1 to the human cellular immune response.

Comparison of 1-D and 2-D LC MS/MS methods for proteomic analysis of human serum

1-D and 2-D LC methods were utilized for proteome analysis of undepleted human serum. Separation of peptides in 2-D LC was performed either with strong cation exchange (SCX)-RP chromatography or with an RP-RP 2-D LC approach. Peptides were identified by MS/MS using a data-independent acquisition approach. A peptide retention prediction model was used to highlight the potential false-positive peptide identifications. When applying selected data filtration, we identified 52 proteins based on 316 peptides in serum in 1-D LC setup. One hundred and eighty-four proteins/1036 peptides and 142 proteins/905 peptides were identified in RP-RP and SCX-RP 2-D LC, respectively. The performance of both 2-D LC methods for proteomic analysis is critically compared.

High-resolution differentiation of transmissible spongiform encephalopathy strains by quantitative N-terminal amino acid profiling (N-TAAP) of PK-digested abnormal prion protein

New forms of transmissible spongiform encephalopathy (TSE) continue to be identified, and consequently sensitive differential diagnosis is increasingly important both for the management of disease in humans and livestock and in providing confidence in the safety of the food chain. TSE diseases are associated with accumulation of protease-resistant prion protein (PrP(Sc)) and detection of this marker protein is central to diagnosis. Proteolysis by proteinase K (PK) generates protease-resistant products (PrP(res)) with partially variable N-termini. The conformation(s) of PrP(Sc) and thus the points of PK cleavage are thought to be dependent on the strain of prion disease. Western blot (WB) analysis of PrP(res) gives characteristic migration patterns that can be used to diagnose TSEs, but the relatively low resolution of this technique limits its ability to differentiate certain disease strains. Mass spectrometry (MS) has the capability to resolve these various PK cleavage sites to the level of individual amino acid residues. In the present study multiple selected reaction monitoring (mSRM) was used to detect and quantify PrP(res) N-terminal tryptic peptides by MS and thus to define the N-terminal amino acid profiles (N-TAAPs) of PrP(res) characteristic for various TSEs in sheep. The fragmentation behaviour of the N-terminal tryptic peptides was studied to allow selection of the transitions specific for each peptide. Different PrP(res) preparation methods were evaluated and the most effective approach applied to differentiate the N-TAAPs corresponding to various sheep TSE isolates. Marked differences were identified between the N-TAAPs of bovine spongiform encephalopathy (BSE) and classical scrapie, and between classical scrapie and the experimental strains SSBP/1 and CH1641, thereby validating this approach as a means of TSE-strain specific diagnosis.

Intrinsic amino acid side-chain hydrophilicity/hydrophobicity coefficients determined by reversed-phase high-performance liquid chromatography of model peptides: comparison with other hydrophilicity/hydrophobicity scales

An accurate determination of the intrinsic hydrophilicity/hydrophobicity of amino acid side-chains in peptides and proteins is fundamental in understanding many area of research, including protein folding and stability, peptide and protein function, protein-protein interactions and peptide/protein oligomerization, as well as the design of protocols for purification and characterization of peptides and proteins. Our definition of intrinsic hydrophilicity/hydrophobicity of side-chains is the maximum possible hydrophilicity/hydrophobicity of side-chains in the absence of any nearest-neighbor effects and/or any conformational effects of the polypeptide chain that prevent full expression of side-chain hydrophilicity/hydrophobicity. In this review, we have compared an experimentally derived intrinsic side-chain hydrophilicity/hydrophobicity scale generated from RP-HPLC retention behavior of de novo designed synthetic model peptides at pH 2 and pH 7 with other RP-HPLC-derived scales, as well as scales generated from classic experimental and calculation-based methods of octanol/water partitioning of Nalpha-acetyl-amino-acid amides or free energy of transfer of free amino acids. Generally poor correlation was found with previous RP-HPLC-derived scales, likely due to the random nature of the peptide mixtures in terms of varying peptide size, conformation and frequency of particular amino acids. In addition, generally poor correlation with the classical approaches served to underline the importance of the presence of a polypeptide backbone when generating intrinsic values. We have shown that the intrinsic scale determined here is in full agreement with the structural characteristics of amino acid side-chains.

Standardization of retention time data for AMT tag proteomics database generation

The combination of liquid chromatography (LC) with mass spectrometry (MS) has become a mainstream proteome analysis strategy. In LC-MS, measured masses possess their "universal" scale derived from atomic mass tables. In contrast, the observed LC retention times (RT) are not tied to a conventional time scale, and depend on experimental conditions. However, RT data, being explicitly orthogonal to MS, offer relevant information for proteome characterization. We present here a strategy for peptides RT data standardization, based on the generation of a standard scale using retention prediction models, which enables sharing of identification databases in the context of multi-laboratory research.

Methods of calculating protein hydrophobicity and their application in developing correlations to predict hydrophobic interaction chromatography retention

Hydrophobic interaction chromatography (HIC) is a key technique for protein separation and purification. Different methodologies to estimate the hydrophobicity of a protein are reviewed, which have been related to the chromatographic behavior of proteins in HIC. These methodologies consider either knowledge of the three-dimensional structure or the amino acid composition of proteins. Despite some restrictions; they have proven to be useful in predicting protein retention time in HIC.

LC-MSsim--a simulation software for liquid chromatography mass spectrometry data

Background

Mass Spectrometry coupled to Liquid Chromatography (LC-MS) is commonly used to analyze the protein content of biological samples in large scale studies. The data resulting from an LC-MS experiment is huge, highly complex and noisy. Accordingly, it has sparked new developments in Bioinformatics, especially in the fields of algorithm development, statistics and software engineering. In a quantitative label-free mass spectrometry experiment, crucial steps are the detection of peptide features in the mass spectra and the alignment of samples by correcting for shifts in retention time. At the moment, it is difficult to compare the plethora of algorithms for these tasks. So far, curated benchmark data exists only for peptide identification algorithms but no data that represents a ground truth for the evaluation of feature detection, alignment and filtering algorithms.

Results

We present LC-MSsim, a simulation software for LC-ESI-MS experiments. It simulates ESI spectra on the MS level. It reads a list of proteins from a FASTA file and digests the protein mixture using a user-defined enzyme. The software creates an LC-MS data set using a predictor for the retention time of the peptides and a model for peak shapes and elution profiles of the mass spectral peaks. Our software also offers the possibility to add contaminants, to change the background noise level and includes a model for the detectability of peptides in mass spectra. After the simulation, LC-MSsim writes the simulated data to mzData, a public XML format. The software also stores the positions (monoisotopic m/z and retention time) and ion counts of the simulated ions in separate files.

Conclusion

LC-MSsim generates simulated LC-MS data sets and incorporates models for peak shapes and contaminations. Algorithm developers can match the results of feature detection and alignment algorithms against the simulated ion lists and meaningful error rates can be computed. We anticipate that LC-MSsim will be useful to the wider community to perform benchmark studies and comparisons between computational tools.

Simulation and Experiment of Temperature and Cosolvent Effects in Reversed Phase Chromatography of Peptides

Experiments and simulations have been carried out for several polar protected peptides in reversed phase chromatography in order to demonstrate how simulation can describe the effects of varying temperature and cosolvent fraction. Comparisons of adsorption chemical potentials from mesoscopic simulations and experimental chromatographic retention data show very good agreement with only one temperature-independent solvent parameter from a single peptide. Such simulations should help guide the design of chromatography experiments with biomolecules and predict retention, including conditions for which empirical correlations such as hydrophobicity scales and molecular descriptors have not been developed.

A robust linear regression based algorithm for automated evaluation of peptide identifications from shotgun proteomics by use of reversed-phase liquid chromatography retention time

BACKGROUND

Rejection of false positive peptide matches in database searches of shotgun proteomic experimental data is highly desirable. Several methods have been developed to use the peptide retention time as to refine and improve peptide identifications from database search algorithms. This report describes the implementation of an automated approach to reduce false positives and validate peptide matches.

RESULTS

A robust linear regression based algorithm was developed to automate the evaluation of peptide identifications obtained from shotgun proteomic experiments. The algorithm scores peptides based on their predicted and observed reversed-phase liquid chromatography retention times. The robust algorithm does not require internal or external peptide standards to train or calibrate the linear regression model used for peptide retention time prediction. The algorithm is generic and can be incorporated into any database search program to perform automated evaluation of the candidate peptide matches based on their retention times. It provides a statistical score for each peptide match based on its retention time.

CONCLUSION

Analysis of peptide matches where the retention time score was included resulted in a significant reduction of false positive matches with little effect on the number of true positives. Overall higher sensitivities and specificities were achieved for database searches carried out with MassMatrix, Mascot and X!Tandem after implementation of the retention time based score algorithm.

Practical implementation of 2D HPLC scheme with accurate peptide retention prediction in both dimensions for high-throughput bottom-up proteomics

We describe the practical implementation of a new RP (pH 10 - pH 2) 2D HPLC-ESI/MS scheme for large-scale bottom-up analysis in proteomics. When compared to the common SCX-RP approach, it provides a higher separation efficiency in the first dimension and increases the number of identified peptides/proteins. We also employed the methodology of our sequence-specific retention calculator (SSRCalc) and developed peptide retention prediction algorithms for both LC dimensions. A diverse set of approximately 10,000 tryptic peptides from the soluble protein fraction of whole NK-type cells gave retention time versus hydrophobicity correlations, with R (2) values of 0.95 for pH 10 and 0.945 for pH 2 (formic acid) separation modes. The superior separation efficiency and the ability to use retention prediction to filter out false-positive MS/MS identifications gives promise that this approach will be a method of choice for large-scale proteomics analyses in the future. Finally, the "semi-orthogonal" separation selectivity permits the concatenation of fractions in the first dimension of separation before the final LC-ESI MS step, effectively cutting the analysis time in half, while resulting in a minimal reduction in protein identification.

Detection of co-eluted peptides using database search methods

BACKGROUND

Current experimental techniques, especially those applying liquid chromatography mass spectrometry, have made high-throughput proteomic studies possible. The increase in throughput however also raises concerns on the accuracy of identification or quantification. Most experimental procedures select in a given MS scan only a few relatively most intense parent ions, each to be fragmented (MS2) separately, and most other minor co-eluted peptides that have similar chromatographic retention times are ignored and their information lost.

RESULTS

We have computationally investigated the possibility of enhancing the information retrieval during a given LC/MS experiment by selecting the two or three most intense parent ions for simultaneous fragmentation. A set of spectra is created via superimposing a number of MS2 spectra, each can be identified by all search methods tested with high confidence, to mimick the spectra of co-eluted peptides. The generated convoluted spectra were used to evaluate the capability of several database search methods - SEQUEST, Mascot, X!Tandem, OMSSA, and RAId_DbS - in identifying true peptides from superimposed spectra of co-eluted peptides. We show that using these simulated spectra, all the database search methods will gain eventually in the number of true peptides identified by using the compound spectra of co-eluted peptides.

OPEN PEER REVIEW

Reviewed by Vlad Petyuk (nominated by Arcady Mushegian), King Jordan and Shamil Sunyaev. For the full reviews, please go to the Reviewers' comments section.

Informatics for peptide retention properties in proteomic LC-MS

Retention times in HPLC yield valuable information for the identification of various analytes and the prediction of peptide retention is useful for the identification of peptides/proteins in LC-MS-based proteomics. Informatics methods such as artificial neural networks and support vector machines capable of solving nonlinear problems made possible the accurate modeling of quantitative structure-retention relationships of peptides (including large polymers) up to 5 kDa to which classical linear models cannot be applied, as well as the proteome-wide prediction of peptide retention. Proteome-wide retention prediction and accurate mass-information facilitate the identification of peptides in complex proteomic samples. In this review, we address recent developments in solid informatics methods and their application to peptide-retention properties in 'bottom-up' shotgun proteomics. We also describe future prospects for the standardization and application of retention times.

Protein abundance profiling of the Escherichia coli cytosol

Background

Knowledge about the abundance of molecular components is an important prerequisite for building quantitative predictive models of cellular behavior. Proteins are central components of these models, since they carry out most of the fundamental processes in the cell. Thus far, protein concentrations have been difficult to measure on a large scale, but proteomic technologies have now advanced to a stage where this information becomes readily accessible.

Results

Here, we describe an experimental scheme to maximize the coverage of proteins identified by mass spectrometry of a complex biological sample. Using a combination of LC-MS/MS approaches with protein and peptide fractionation steps we identified 1103 proteins from the cytosolic fraction of the Escherichia coli strain MC4100. A measure of abundance is presented for each of the identified proteins, based on the recently developed emPAI approach which takes into account the number of sequenced peptides per protein. The values of abundance are within a broad range and accurately reflect independently measured copy numbers per cell.

As expected, the most abundant proteins were those involved in protein synthesis, most notably ribosomal proteins. Proteins involved in energy metabolism as well as those with binding function were also found in high copy number while proteins annotated with the terms metabolism, transcription, transport, and cellular organization were rare. The barrel-sandwich fold was found to be the structural fold with the highest abundance. Highly abundant proteins are predicted to be less prone to aggregation based on their length, pI values, and occurrence patterns of hydrophobic stretches. We also find that abundant proteins tend to be predominantly essential. Additionally we observe a significant correlation between protein and mRNA abundance in E. coli cells.

Conclusion

Abundance measurements for more than 1000 E. coli proteins presented in this work represent the most complete study of protein abundance in a bacterial cell so far. We show significant associations between the abundance of a protein and its properties and functions in the cell. In this way, we provide both data and novel insights into the role of protein concentration in this model organism.

Statistical learning of peptide retention behavior in chromatographic separations: a new kernel-based approach for computational proteomics

Background

High-throughput peptide and protein identification technologies have benefited tremendously from strategies based on tandem mass spectrometry (MS/MS) in combination with database searching algorithms. A major problem with existing methods lies within the significant number of false positive and false negative annotations. So far, standard algorithms for protein identification do not use the information gained from separation processes usually involved in peptide analysis, such as retention time information, which are readily available from chromatographic separation of the sample. Identification can thus be improved by comparing measured retention times to predicted retention times. Current prediction models are derived from a set of measured test analytes but they usually require large amounts of training data.

Results

We introduce a new kernel function which can be applied in combination with support vector machines to a wide range of computational proteomics problems. We show the performance of this new approach by applying it to the prediction of peptide adsorption/elution behavior in strong anion-exchange solid-phase extraction (SAX-SPE) and ion-pair reversed-phase high-performance liquid chromatography (IP-RP-HPLC). Furthermore, the predicted retention times are used to improve spectrum identifications by a p-value-based filtering approach. The approach was tested on a number of different datasets and shows excellent performance while requiring only very small training sets (about 40 peptides instead of thousands). Using the retention time predictor in our retention time filter improves the fraction of correctly identified peptide mass spectra significantly.

Conclusion

The proposed kernel function is well-suited for the prediction of chromatographic separation in computational proteomics and requires only a limited amount of training data. The performance of this new method is demonstrated by applying it to peptide retention time prediction in IP-RP-HPLC and prediction of peptide sample fractionation in SAX-SPE. Finally, we incorporate the predicted chromatographic behavior in a p-value based filter to improve peptide identifications based on liquid chromatography-tandem mass spectrometry.

Analysis of protein three-dimension structure using amino acids depths

The issue of amino acid depth in proteins gives important insights to our understanding of protein's three-dimensional structure. There has already been much research done in mathematical and statistical sciences regarding the general definitions, properties and algorithms describing the particle depth of spatially extended systems. We constructed a method of calculating the amino acids depths and applied it to a set of 527 protein structures. We propose the introduction of amino acid depth tendency factors for three-dimensional structures of proteins. The depth tendency factors relate not only to the hydrophobicity indices but also to the electrostatic charge. We found a relationship between the protein size and the number of residues using the distance between the deepest residue and surface residues. We made a prediction regarding the number of residues on the surface of a protein, the deepest amino acid, and the average depth, all of which are fitted well to a linear functional relationship with the length of the protein. Finally, we have predicted the depths of multiple peptides in protein's three-dimension structure.

A statistical learning approach to the modeling of chromatographic retention of oligonucleotides incorporating sequence and secondary structure data

We propose a new model for predicting the retention time of oligonucleotides. The model is based on nu support vector regression using features derived from base sequence and predicted secondary structure of oligonucleotides. Because of the secondary structure information, the model is applicable even at relatively low temperatures where the secondary structure is not suppressed by thermal denaturing. This makes the prediction of oligonucleotide retention time for arbitrary temperatures possible, provided that the target temperature lies within the temperature range of the training data. We describe different possibilities of feature calculation from base sequence and secondary structure, present the results and compare our model to existing models.

On-line coupling of SPE and CE-MS for peptide analysis

An on-line SPE-CE-MS system has been developed for the analysis of peptides. Analytes are preconcentrated using a C(18) microcolumn (5 x 0.5 mm id), and then introduced into the CE system via a valve interface. The CE system with a Polybrene-poly(vinylsulfonate) bilayer coated capillary is combined with an ion-trap mass spectrometer via ESI using a coaxial sheath-liquid sprayer. The on-line coupling of the SPE and CE step by the valve interface is advantageous because it allows an independent functioning of the system parts. Optimization of the SPE-CE system was performed using UV detection. Subsequently, the SPE-CE system has been coupled to the ion-trap mass spectrometer. Test solutions with enkephalin peptides (50 ng/mL) were used for evaluation of system performance. Repeatability of effective mobility and peak area ratio of the two enkephalins were within 1.2% and 9% RSD, respectively. The analysis of 1:1 v/v diluted cerebrospinal fluid samples spiked with enkephalin peptides showed detection limits (S/N = 3) in the range of 1.5-3 ng/mL (around 5 nM), which were similar to those obtained for enkephalin test solutions. Moreover, the potential of the on-line SPE-CE-MS system was demonstrated by the analysis of a cytochrome C digest. Some hydrophilic peptides did not show sufficient retention on the SPE column, and were lost during preconcentration. Nonetheless, positive identification of the protein was achieved, indicating the feasibility of the system for proteomics.

Peptide retention prediction applied to proteomic data analysis

A retention prediction model was developed for peptides separated in reversed-phase chromatography. The model was utilized to identify and exclude the false positive (FP) peptide identifications obtained via database search. The selected database included human proteins, as well as decoy sequences of random proteins. The FP peptide detection rate was defined either as number of retention time outliers, or random decoy sequence identifications. The FP rate for various MASCOT scores was calculated. The peptides identified in one-dimensional (1D) and two-dimensional (2D) liquid chromatography/mass spectrometry (LC/MS) experiments were validated by prediction models. Multi-dimensional LC was based on two orthogonal reversed-phase chromatography modes; prediction models were successfully applied for data filtering in both separation dimensions.

Efficient enrichment and desalting of protein digests using magnetic mesocellular carbon foams in matrix-assisted laser desorption/ionization mass spectrometry

We demonstrate that magnetic mesocellular carbon foams (Mag-MCF-C) can be effectively used for enrichment and desalting of protein digests or peptides in matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). The large mesocellular pores and surface area of Mag-MCF-C are likely to mainly contribute to high efficiency in enrichment and desalting of protein digests. The magnetic property of Mag-MCF-C enabled easy and simple enrichment and desalting process comprising adsorption, washing, and separation steps by using an external magnet. Following elution from Mag-MCF-C by using a matrix solution (CHCA in 70% ACN/0.1% TFA), the peptides were subjected to MALDI-MS analysis. As a result, MALDI mass spectra of peptides or tryptic protein digests were distinct even at a peptide concentration as low as 50 pM. The use of Mag-MCF-C resulted in significantly improved sequence coverage for protein identification when compared to other conventional methods. Mag-MCF-C will find applications in mass spectrometric analysis of low abundance peptides or protein digests with high sensitivity.

Requirements for prediction of peptide retention time in reversed-phase high-performance liquid chromatography: hydrophilicity/hydrophobicity of side-chains at the N- and C-termini of peptides are dramatically affected by the end-groups and location

The value of reversed-phase high-performance liquid chromatography (RP-HPLC) and the field of proteomics would be greatly enhanced by accurate prediction of retention times of peptides of known composition. The present study investigates the hydrophilicity/hydrophobicity of amino acid side-chains at the N- and C-termini of peptides while varying the functional end-groups at the termini. We substituted all 20 naturally occurring amino acids at the N- and C-termini of a model peptide sequence, where the functional end-groups were N(alpha)-acetyl-X- and N(alpha)-amino-X- at the N-terminus and -X-C(alpha)-carboxyl and -X-C(alpha)-amide at the C-terminus. Amino acid coefficients were subsequently derived from the RP-HPLC retention behaviour of these peptides and compared to each other as well as to coefficients determined in the centre of the peptide chain (internal coefficients). Coefficients generated from residues substituted at the C-terminus differed most (between the -X-C(alpha)-carboxyl and -X-C(alpha)-amide peptide series) for hydrophobic side-chains. A similar result was seen for the N(alpha)-acetyl-X- and N(alpha)-amino-X- peptide series, where the largest differences in coefficient values were observed for hydrophobic side-chains. Coefficients derived from substitutions at the C-terminus for hydrophobic amino acids were dramatically different compared to internal coefficients for hydrophobic side-chains, ranging from 17.1 min for Trp to 4.8 min for Cys. In contrast, coefficients derived from substitutions at the N-terminus showed relatively small differences from the internal coefficients. Subsequent prediction of peptide retention time, within an error of just 0.4 min, was achieved by a predictive algorithm using a combination of internal coefficients and coefficients for the C-terminal residues. For prediction of peptide retention time, the sum of the coefficients must include internal and terminal coefficients.