Peptide Retention Time Research Articles

Peptide retention time prediction for electrostatic repulsion-hydrophilic interaction chromatography

Electrostatic Repulsion-Hydrophilic Interaction Chromatography (ERLIC) is one of the legacy separation tools developed by Dr. Andrew Alpert and has been used for developing unique separation methods of hydrophilic compounds, including peptides. In the past it has been studied using designed peptide libraries to elucidate major features of its separation mechanism, while comprehensive peptide retention modeling for ERLIC is still lacking. In this work we employed a proteomics-derived ∼170,000 peptide retention datasets to evaluate major ERLIC retention features using the framework of our Sequence-Specific Retention Calculator model. The separation conditions were adjusted to obtain a wider proteome coverage, particularly for non-modified peptides, resulting in a superior separation orthogonality for a 2D LC combination with reversed-phase C18 LC-MS in the second dimension. The SSRCalc ERLIC model presents a consistent theme with the existing ERLIC retention mechanism, reflecting a dependence on peptide orientation and the position of charged and hydrophilic residues across the peptide backbone. R2 values of 0.935 and 0.955 accuracy were demonstrated for the standard interpretable SSRCalc model and machine learning algorithm, respectively. The effects of various PTMs on peptide retention were evaluated in this study, covering spontaneous (oxidation, deamidation) and enzymatic (N-terminal acetylation, phosphorylation, glycosylation) modifications.

Peptide retention time prediction for hydrophilic interaction liquid chromatography at acidic pH in formic-acid based eluents

Peptide separation selectivity was evaluated for hydrophilic interaction liquid chromatography (HILIC) ZIC-HILIC, ZIC-cHILIC, and XBridge Amide sorbents using formic acid as eluent additive (pH 2.7). Sequence-specific retention prediction algorithms were trained using retention datasets of ∼30,000 peptides for each column. Our retention models were able to attain ∼0.98 R2-value and yielded retention coefficients that can be probed to understand peptide-stationary phase interaction. Overall, the hydrophilicity for these columns decreased when the mobile phase changed pH from 4.5 to 2.7, when using 0.1 % formic acid in the mobile phase. The acidic residues became protonated, and the resultant hydrophilic interaction is dampened at the lower pH, leaving only the basic residues as the primary hydrophilic interactors. Hence, peptides of increasing charge have higher retention. In this comparison between the three columns, ZIC-HILIC has the highest chromatographic resolution between groups of peptides of different charge. From the position-dependent retention coefficients for ZIC-HILIC at pH 2.7, we found that the amino acids at the terminal positions of the peptide modulate the basicity of the N-terminal amino group or the C-terminal Arg/Lys for tryptic peptides. With respect to the separation orthogonality between HILIC and acidic pH RPLC for two dimensional separations, the orthogonality values were lower at pH 2.7 than operating HILIC at pH 4.5 for the first dimension. We also demonstrate that ZIC-HILIC was able to distinguish citrullinated and deamidated peptides based on predicted retention values.

Evaluating C18 stationary phases with a positively charged surface for proteomic LC-MS applications using mobile phase acidified with reduced formic acid concentration

We have recently demonstrated the ability of a C18 stationary phase with a positively charged surface (PCS-C18) to provide superior chromatographic separation of peptides using mobile phase acidified with a mere 0.01 % formic acid, significantly improving MS sensitivity. Here, we examined three columns packed with different PCS-C18 phases using the MS-favorable mobile phase acidified with low formic acid concentrations to establish the impact of separation performance and better MS sensitivity on peptide identifications. The surface charge interaction was evaluated using the retention of nitrate. The highest interaction was observed for the AdvanceBio Peptide Plus column. A surface charge-dependent shift in the retention time of peptides was confirmed with a change in formic acid concentration in the mobile phase. The separation performance of the columns with MS-favorable mobile phase acidified with low concentrations of formic acid was evaluated using well-characterized peptides. The loading capacity was assessed using a basic peptide with three lysine residues. Good chromatographic peak shapes and high loading capacity were observed for the Acquity Premier CSH C18 column, even when using a mobile phase acidified with 0.01 % formic acid. The extent of improvement in peptide identification when using reduced formic acid concentration was evaluated by analyzing the tryptic digest of trastuzumab and tryptic digest of whole bacteria cell lysate. Each column provided improved MS signal intensity and peptide identification when using the mobile phase with 0.01 % formic acid. The ability of the Acquity Premier CSH C18 column to provide better separation of peptides, even with a reduced formic acid concentration in the mobile phase, boosted MS signal intensity by 65 % and increased the number of identified peptides from the bacterial sample by 19 %. Our study confirms that significant improvement in the proteomic outputs can be achieved without additional costs only by tailoring the chemistry of the stationary phase to the composition of the mobile phase. Our results can help researchers understand the retention mechanism of peptides on the PCS-C18 stationary phases using low-ionic strength mobile phases and, more importantly, select the best-suited stationary phases for their LC-MS proteomic applications.

Calib-RT: an open source python package for peptide retention time calibration in DIA mass spectrometry data.

The data independent acquisition (DIA) mass spectrometry (MS) method is increasingly popular in the field of proteomics. But the loss of the correspondence between peptide ions and their spectra in DIA makes the identification challenging. One effective approach to reduce false positive identification is to calculate the deviation between the peptide's estimated retention time (RT) and measured RT. During this process, scaling the spectral library RT into the estimated RT, known as the RT calibration, is a prerequisite for calculating the deviation. Currently, within the DIA algorithm ecosystem, there is a lack of engine-independent and readily usable RT calibration toolkits. In this work, we introduce Calib-RT, a RT calibration method tailored to the characteristics of RT data. This method can achieve the nonlinear calibration across various data scales and tolerate a certain level of noise interference. Calib-RT is expected to enrich the open source DIA algorithm toolchain and assist in the development of DIA identification algorithms. Calib-RT is released as an open source software under the MIT license and can be installed from PyPi as a python module. The source code is available on GitHub at https://github.com/chenghui03/Calib_RT.

Solving the retention time repeatability problem of hydrophilic interaction liquid chromatography

Hydrophilic interaction (liquid) chromatography (HILIC) has become the first choice LC mode for the separation of hydrophilic analytes. Numerous studies reported the poor retention time repeatability of HILIC. The problem was often ascribed to slow equilibration and insufficient re-equilibration time to establish the sensitive semi-immobilized water layer at the interface of the polar stationary phase and the bulk mobile phase. In this study, we compare retention time repeatability in HILIC for borosilicate glass and PFA (co-polymer of tetrafluoroethylene and perfluoroalkoxyethylene) solvent bottles. During this study, we observed peak patterns shifting towards higher retention times (for metabolites and peptides) and lower retention times (oligonucleotide sample) with ongoing analysis time when standard borosilicate glass bottles were used as solvent reservoirs. It was hypothesized that release of ions (sodium, potassium, borate, etc.) from the borosilicate glass bottles leads to alterations (thickness and electrostatic screening effects) in the semi-immobilized water layer which is adsorbed to the polar stationary phase surface under acetonitrile-rich eluents in HILIC with concomitant shifts in retention. When PFA solvent bottles were employed instead of borosilicate glass, retention time repeatability was greatly improved and changed from average 8.4 % RSD for the tested metabolites with borosilicate glass bottles to 0.14 % RSD for the PFA solvent bottles (30 injections over 12 h). Similar improvements were observed for peptides and oligonucleotides. This simple solution to the retention time repeatability problem in HILIC might contribute to a better acceptance of HILIC, especially in fields like targeted and untargeted metabolomics, peptide and oligonucleotide analysis.

Identification of horn-derived animal medicines based on detection of keratin-derived characteristic peptides

Due to the highly stable structure of keratin, the extraction and dissolution steps of animal medicines rich in keratin are complex, which seriously restricts the detection efficiency and flux. Therefore, this study simplified the pre-treatment steps of horn samples and optimized the detection methods of characteristic peptides to improve the efficiency of identifying the specificity of horn-derived animal medicines. For detection of the characteristic peptides in horn-derived animal medicines treated with/without iodoace-tamide(IAA), the ion pair conditions of the characteristic peptides were optimized, and the retention time, intensity and other data of the specific peptides were compared between the samples treated with/without IAA. Two pre-treatment methods, direct enzymatic hydrolysis and total protein extraction followed by enzymatic hydrolysis, were used to prepare horn-derived animal medicine samples. The effects of different methods on the detection of specific peptides in the samples of Saiga antelope horn, water buffalo horn, goat horn, and yak horn were compared regarding the retention time of specific peptides and ion intensity. The results indicated that after direct enzymatic hydrolysis, the specific peptides in the samples without IAA treatment can be detected. Compared with the characteristic peptides in the samples treated with IAA, their retention time shifted back and the mass spectrometry response slightly decreased. The specific peptides of the samples without IAA treatment had good specificity and did not affect the specificity identification of horn-derived animal medicines. Overall, the process of direct enzymatic hydrolysis can be used to treat horn samples, omitting the steps of protein extraction and dithiothreitol and IAA treatment, significantly improving the pre-treatment efficiency without affecting the specificity identification of horn-derived animal medicines. This study provides ideas for quality research and standard improvement of horn-derived animal medicines.

Intensity and retention time prediction improves the rescoring of protein-nucleic acid cross-links.

In protein-RNA cross-linking mass spectrometry, UV or chemical cross-linking introduces stable bonds between amino acids and nucleic acids in protein-RNA complexes that are then analyzed and detected in mass spectra. This analytical tool delivers valuable information about RNA-protein interactions and RNA docking sites in proteins, both in vitro and in vivo. The identification of cross-linked peptides with oligonucleotides of different length leads to a combinatorial increase in search space. We demonstrate that the peptide retention time prediction tasks can be transferred to the task of cross-linked peptide retention time prediction using a simple amino acid composition encoding, yielding improved identification rates when the prediction error is included in rescoring. For the more challenging task of including fragment intensity prediction of cross-linked peptides in the rescoring, we obtain, on average, a similar improvement. Further improvement in the encoding and fine-tuning of retention time and intensity prediction models might lead to further gains, and merit further research.

Toward an Ultimate Solution for Peptide Retention Time Prediction: The Effect of Column Temperature on Separation Selectivity.

We studied the effect of the column temperature on the selectivity of reversed-phase peptide separation in bottom-up proteomics. The number of peptide identifications from 2 h liquid chromatography with tandem mass spectrometry (LC-MS/MS) acquisitions reaches a plateau at 45-55 °C, driven simultaneously by improved separation efficiency, a gradual decrease in peptide retention, and possible on-column degradation of peptides at elevated temperatures. Performing 2D LC-MS/MS acquisitions at 25, 35, 45, and 55 °C resulted in the identification of ∼100,000 and ∼120,000 unique peptides for nonmodified and tandem mass tags (TMT)-labeled samples, respectively. These peptide collections were used to investigate the temperature-driven retention features. The latter is governed by the specific temperature response of individual residues, peptide hydrophobicity and length, and amphipathic helicity. On average, peptide retention decreased by 0.56 and 0.5% acetonitrile for each 10 °C increase for label-free and TMT-labeled peptides, respectively. This generally linear response of retention shifts allowed the extrapolation of predictive models beyond the studied temperature range. Thus, (trap) column cooling from room temperature to 0 °C will allow the retention of an additional 3% of detectable tryptic peptides. Meanwhile, the application of 90 °C would result in the loss of ∼20% of tryptic peptides that were amenable to MS/MS-based identification.

MARS an improved de novo peptide candidate selection method for non-canonical antigen target discovery in cancer

Understanding the nature and extent of non-canonical human leukocyte antigen (HLA) presentation in tumour cells is a priority for target antigen discovery for the development of next generation immunotherapies in cancer. We here employ a de novo mass spectrometric sequencing approach with a refined, MHC-centric analysis strategy to detect non-canonical MHC-associated peptides specific to cancer without any prior knowledge of the target sequence from genomic or RNA sequencing data. Our strategy integrates MHC binding rank, Average local confidence scores, and peptide Retention time prediction for improved de novo candidate Selection; culminating in the machine learning model MARS. We benchmark our model on a large synthetic peptide library dataset and reanalysis of a published dataset of high-quality non-canonical MHC-associated peptide identifications in human cancer. We achieve almost 2-fold improvement for high quality spectral assignments in comparison to de novo sequencing alone with an estimated accuracy of above 85.7% when integrated with a stepwise peptide sequence mapping strategy. Finally, we utilize MARS to detect and validate lncRNA-derived peptides in human cervical tumour resections, demonstrating its suitability to discover novel, immunogenic, non-canonical peptide sequences in primary tumour tissue.

Radiolabeled NGR-Based Heterodimers for Angiogenesis Imaging: A Review of Preclinical Studies.

Since angiogenesis/neoangiogenesis has a major role in tumor development, progression and metastatic spread, the establishment of angiogenesis-targeting imaging and therapeutic vectors is of utmost significance. Aminopeptidase N (APN/CD13) is a pivotal biomarker of angiogenic processes abundantly expressed on the cell surface of active vascular endothelial and various neoplastic cells, constituting a valuable target for cancer diagnostics and therapy. Since the asparagine-glycine-arginine (NGR) sequence has been shown to colocalize with APN/CD13, the research interest in NGR-peptide-mediated vascular targeting is steadily growing. Earlier preclinical experiments have already demonstrated the imaging and therapeutic feasibility of NGR-based probes labeled with different positron emission tomography (PET) and single-photon emission computed tomography (SPECT) radionuclides, including Gallium-68 (68Ga), Copper-64 (64Cu), Technetium-99m (99mTc), Lutetium-177 (177Lu), Rhenium-188 (188Re) or Bismuth-213 (213Bi). To improve the tumor binding affinity and the retention time of single-receptor targeting peptides, NGR motifs containing heterodimers have been introduced to identify multi-receptor overexpressing malignancies. Preclinical studies with various tumor-bearing experimental animals provide useful tools for the investigation of the in vivo imaging behavior of NGR-based heterobivalent ligands. Herein, we review the reported preclinical achievements on NGR heterodimers that could be highly relevant for the development of further target-specific multivalent compounds in diagnostic and therapeutic settings.

Assessing the hydrophobicity of glycopeptides using reversed-phase liquid chromatography and tandem mass spectrometry

Retention time is one of the most important parameters that has been widely used to demonstrate the separation results obtained from liquid chromatography (LC) platforms. However, retention time can shift when samples are tested with different instruments and laboratories, which hinders the identification process of analytes when comparing data collected from different LC systems. To address this problem, hydrophobicity index was introduced for retention time normalization of the glycopeptides separated by reversed-phase LC (RPLC). Tandem MS was used for the detection and identification of glycopeptides. In addition, the influence of different types of glycans on the hydrophobicity of peptide backbones was studied by comparing the retention time of glycopeptides with their non-glycosylated counterparts. The hydrophobicity of tryptic digested glycopeptides derived from model glycoproteins, including bovine fetuin, α1-acid glycoprotein, and haptoglobin from human plasma, were evaluated based on the hydrophobicity index of the standard peptides from a peptide retention time calibration mixture. The reduction of hydrophobicity of multiple peptide backbones was observed due to the hydrophilic glycan structures. By comparing the hydrophobicity index of glycopeptides collected from different time and instruments, the day-to-day and lab-to-lab comparisons suggested high reliability and reproducibility of this approach. The RSD% of hydrophobicity index from inter-lab experiments was 1.2%, while the RSD% of retention time was 5.1%. Then, the applications of this method were demonstrated on complex glycopeptide samples extracted from human blood serum. The hydrophobicity index can be applied to address the retention time shift when using different instruments, thereby boosting confidence of the characterization of glycopeptides.

Immunoprecipitation methods impact the peptide repertoire in immunopeptidomics.

Mass spectrometry-based immunopeptidomics is the only unbiased method to identify naturally presented HLA ligands, which is an indispensable prerequisite for characterizing novel tumor antigens for immunotherapeutic approaches. In recent years, improvements based on devices and methodology have been made to optimize sensitivity and throughput in immunopeptidomics. However, developments in ligand isolation, mass spectrometric analysis, and subsequent data processing can have a marked impact on the quality and quantity of immunopeptidomics data. In this work, we compared the immunopeptidome composition in terms of peptide yields, spectra quality, hydrophobicity, retention time, and immunogenicity of two established immunoprecipitation methods (column-based and 96-well-based) using cell lines as well as primary solid and hematological tumor samples. Although, we identified comparable overall peptide yields, large proportions of method-exclusive peptides were detected with significantly higher hydrophobicity for the column-based method with potential implications for the identification of immunogenic tumor antigens. We showed that column preparation does not lose hydrophilic peptides in the hydrophilic washing step. In contrast, an additional 50% acetonitrile elution could partially regain lost hydrophobic peptides during 96-well preparation, suggesting a reduction of the bias towards the column-based method but not completely equalizing it. Together, this work showed how different immunoprecipitation methods and their adaptions can impact the peptide repertoire of immunopeptidomic analysis and therefore the identification of potential tumor-associated antigens.

Statistical methods and its applications in Whey protein binding (Bos Taurus)

Milk is made up of two proteins, such as casein (80%) and whey (20%). These proteins, which include membrane globular proteins, glycoproteins, and lipoproteins, are mostly found in the outer layer of milk fat globule membranes. WP can be separated from the casein in milk or formed as a byproduct of cheese making and contains all nine essential amino acids with a low lactose content. Whey protein has many health benefits, including improving strength and body composition, promoting weight loss, boosting metabolism, stifling hunger, maintaining good health, avoiding nourishing the bones, reducing inflammation, promoting quick wound healing, lowering blood pressure, repairing and preserving muscle tissues etc. Many dairy and food-based industries eagerly explore the accurate retention time for the formation of health related new byproducts. As of date, there are very few studies available in the world literature citation database and too many micro and macro studies on WP are currently being undertaken globally, they provide insight into the binding of particular amino acids (Seq. peptides) and their accurate lag time of binding to the formation of "helixes and sheaths. In this paradigm, statistical tool are very important to determine average time and WP algorithms for analysis of seq amino acids and the metabolic pathway's action. Overall, the present research attempts to extrapolate the Seq peptide's binding capacity and its retention time for amino acids in order to fill the aforementioned research gap. For the model demonstration we used the commonly cited whey protein peptide (Seq) NCBS, which was considered to prove the hypothesis. The Markova two state random and Monte Carlo simulation (superimposed artificial neural network) models were formulated to extract transient probabilities and retention times of Seq peptides. As per the results, the binding capacity was found to be excellent with a good retention time (5 to 11.50 minutes) in cross-linked and unlinked cases (10 to 13.50 minutes). Bioinformatics, the dairy industry, and life scientists will find these more recent discoveries to be very helpful for measuring likelihoods and transient probabilities of binding.

Characterization of the natural peptidome of four leeches by integrated proteogenomics and pseudotargeted peptidomics

Animal-derived drugs are an indispensable part of folk medicine worldwide. However, their chemical constituents are poorly approached, which leads to the low level of the quality standard system of animal-derived drugs and further causes a chaotic market. Natural peptides are ubiquitous throughout the organism, especially in animal-derived drugs. Thus, in this study, we used multi-source leeches, including Hirudo nipponica (HN), Whitmania pigra (WP), Whitmania acranulata (WA), and Poecilobdella manillensis (PM), as a model. A strategy integrating proteogenomics and novel pseudotargeted peptidomics was developed to characterize the natural peptide phenotype and screen for signature peptides of four leech species. First, natural peptides were sequenced against an in-house annotated protein database of closely related species constructed from RNA-seq data from the Sequence Read Archive (SRA) website, which is an open-sourced public archive resource. Second, a novel pseudotargeted peptidomics integrating peptide ion pair extraction and retention time transfer was established to achieve high coverage and quantitative accuracy of the natural peptides and to screen for signature peptides for species authentication. In all, 2323 natural peptides were identified from four leech species whose databases were poorly annotated. The strategy was shown to significantly improve peptide identification. In addition, 36 of 167 differential peptides screened by pseudotargeted proteomics were identified, and about one-third of them came from the leucine-rich repeat domain (LRR) proteins, which are widely distributed in organisms. Furthermore, six signature peptides were screened with good specificity and stability, and four of them were validated by synthetic standards. Finally, a dynamic multiple reaction monitoring (dMRM) method based on these signature peptides was established and revealed that one-half of the commercial samples and all of the Tongxinluo capsules were derived from WP. All in all, the strategy developed in this study was effective for natural peptide characterization and signature peptide screening, which could also be applied to other animal-derived drugs, especially for modelless species that are less studied in protein database annotation.

Preparation of open tubular capillary column covalently coated with polystyrene sulfonate with 4,4'-azobis (4-cyanopentanoyl chloride) as polymerization initiator for electrochromatographic separation of alkaloids, sulfonamides and peptides.

An open tubular capillary electrochromatography column covalently bonded with polystyrene sulphonate was prepared via in situ polymerization using functionalized azo-initiator 4,4'-Azobis(4-cyanopentanoyl chloride). Scanning electron, fluorescence and atomic force microscopy techniques showed the formation of a relatively rough layer of polymer. In addition, -CN and C = O stretching vibrations from infrared spectroscopy proved the successful immobilization of the azo-initiator through covalent bonding and X-rayPhotoelectronSpectroscopy confirmed the elemental composition of the formed polymer layer. The prepared column was found to be appropriate for small and medium-sized molecules separation. Compared to bare fused silica capillary column higher selectivity and resolution were obtained for the separation of alkaloids, sulfonamides and peptides as a result of the electrostatic and pi-pi stacking interactions between the small organic molecules and the coated column without compromising the electroosmotic flow mobility. Separation efficiency was also increased compared to the bare capillary for the separation of alkaloids (about 1.5 times). Moreover, intraday, inter day, intra batch and inter batch relative standard deviation values of retention time and peak area of peptides were within 2% and 10%, respectively, indicating a good repeatability of the column preparation procedure. The developed method for the covalent bonding of polymers through a functionalized azo-initiator could represent a promising stable method for the preparation of open tubular column. This article is protected by copyright. All rights reserved.

Quantitative PET tracking of intra-articularly administered 89Zr-peptide-decorated nanoemulsions.

Intra-articular (IA) administration of drugs for the treatment of diseases such as rheumatoid arthritis, osteoarthritis and psoriatic arthritis is a common strategy; however, the rapid clearance from the synovial fluid restricts their effectivity due to the limited retention time. Drug Delivery Systems (DDS) are currently being developed to increase their joint retention time. This study compares the biodistribution and retention time of a senolytic peptide (PEP), with potential application in osteoarthritis disease, and this senolytic peptide encapsulated in a DDS based on a lipid nanoemulsion (PEPNE) by using positron emission tomography (PET) imaging. To this aim, the PEP was conjugated with a chelating agent (DFO) and radiolabeled with zirconium-89 (89Zr). Then, [89Zr]-PEP was encapsulated in a novel nanoemulsion formulation, composed by vitamin E, sphingomyelin, and a lipid-PEG. Afterward, healthy rats were administered with either the [89Zr]-PEP or the [89Zr]-PEP-NE via IA injection and underwent PET scans at 0.5-, 24-, 48-, 72-, 168-, 240- and 336h post-injection. To assess the biodistribution of both radiotracers, several volume-of-interest were manually drawn in different organs of the rat body and the %ID/organ was calculated. The [89Zr]-PEP was successfully encapsulated in the NE and their physicochemical properties were minimally affected by the radiolabeling buffer. Adequate stability of both [89Zr]-PEP and [89Zr]-PEP-NE was found in synovial fluid over 72h. Quantitative data from PET images revealed a significantly higher [89Zr]-PEP-NE retention in the injected knee than with [89Zr]-PEP in all follow-up PET scans. The [89Zr]-PEP %ID/organ values in the liver and kidney were significantly higher than those from [89Zr]-PEP-NE, which might indicate a faster elimination of the [89Zr]-PEP. Therefore, the study highlights the higher retention time on the target site of the [89Zr]-PEP-NE which may improve the therapeutic effects of the peptide. Thereby, the novel nanoemulsion formulation seems to be a successful DDS for IA injection. In addition, these results represent the first study that evaluates the distribution of a PET-guided DDS after its IA administration.

Terminally Phosphorylated Triblock Polyethers Acting Both as Templates and Pore-Forming Agents for Surface Molecular Imprinting of Monoliths Targeting Phosphopeptides.

The novel process reported here described the manufacture of monolithic molecularly imprinted polymers (MIPs) using a terminally functionalized block copolymer as the imprinting template and pore-forming agent. The MIPs were prepared through a step-growth polymerization process using a melamine-formaldehyde precondensate in a biphasic solvent system. Despite having a relatively low imprinting factor, the use of MIP monolith in liquid chromatography demonstrated the ability to selectively target desired analytes. An MIP capillary column was able to separate monophosphorylated peptides from a tryptic digest of bovine serum albumin. Multivariate data analysis and modeling of the phosphorylated and nonphosphorylated peptide retention times revealed that the number of phosphorylations was the strongest retention contributor for peptide retention on the monolithic MIP capillary column.

Exploring the variable space of shallow machine learning models for reversed-phase retention time prediction

Peptide retention time (RT) prediction algorithms are tools to study and identify the physicochemical properties that drive the peptide-sorbent interaction. Traditional RT algorithms use multiple linear regression with manually curated parameters to determine the degree of direct contribution for each parameter and improvements to RT prediction accuracies relied on superior feature engineering. Deep learning led to a significant increase in RT prediction accuracy and automated feature engineering via chaining multiple learning modules. However, the significance and the identity of these extracted variables are not well understood due to the inherent complexity when interpreting “relationships-of-relationships” found in deep learning variables. To achieve both accuracy and interpretability simultaneously, we isolated individual modules used in deep learning and the isolated modules are the shallow learners employed for RT prediction in this work. Using a shallow convolutional neural network (CNN) and gated recurrent unit (GRU), we find that the spatial features obtained via the CNN correlate with real-world physicochemical properties namely cross-collisional sections (CCS) and variations of assessable surface area (ASA). Furthermore, we determined that the discovered parameters are “micro-coefficients” that contribute to the “macro-coefficient” – hydrophobicity. Manually embedding CCS and the variations of ASA to the GRU model yielded an R2 = 0.981 using only 525 variables and can represent 88% of the ∼110,000 tryptic peptides used in our dataset. This work highlights the feature discovery process of our shallow learners can achieve beyond traditional RT models in performance and have better interpretability when compared with the deep learning RT algorithms found in the literature.

Offline Peptide Fractionation and Parallel Reaction Monitoring MS for the Quantitation of Low-Abundance Plasma Proteins.

Mass spectrometry (MS)-based protein quantitation is an attractive means for research and diagnostics due to its high specificity, precision, sensitivity, versatility, and the ability to develop multiplexed assays for the "absolute" quantitation of virtually any protein target. However, due to the large dynamic range of protein concentrations in blood, high abundance proteins in blood plasma hinder the detectability and quantification of lower-abundance proteins which are often relevant in the context of different diseases. Here we outline a streamlined method involving offline high-pH reversed-phase fractionation of human plasma samples followed by the quantitative analysis of specific fractions using nanoLC-parallel reaction monitoring (PRM) on a Q Exactive Plus mass spectrometer for peptide detection and quantitation with increased sensitivity. Because we use a set of synthetic peptide standards, we can more efficiently determine the precise retention times of the target peptides in the first-dimensional separation and specifically collect eluting fractions of interest for the subsequent targeted MS quantitation, making the analysis faster and easier. An eight-point standard curve was generated by serial dilution of a mixture of previously validated unlabeled ("light") synthetic peptides of interest at known concentrations. The corresponding heavy stable-isotope-labeled standard (SIS) analogues were used as normalizers to account for losses during sample processing and analysis. Using this method, we were able to improve the sensitivity of plasma protein quantitation by up to 50-fold compared to using nanoLC-PRM alone.

A Program for Predicting the Retention Time of Peptides with Post-Translational Modifications

This paper describes the Retention Time Predictor (RTP) program and web service for predicting the retention time of peptides on a chromatographic column in mass spectrometry experiments. Taking into account post-translational modifications of peptides the program represents a modification of the well-known SSRCalc version 3 (Krokhin, Anal. Chem. 2006, 78(22), 7785-7795). The values of retention coefficients for modified amino acid residues and the algorithm for calculating the isoelectric point value were from the pIPredict program (Skvortsov et al., Biomed. Chem. Res. Meth. 2021, 4(4), e00161). Modifications described in the program include (i) Tandem Mass Tag (TMT) and Isobaric Tags for Relative and Absolute Quantification (iTRAQ) labels; (ii) acetylation, formylation, and methylation of the N-terminal residue and/or lysine side chain; (iii) carbamidomethylation of cysteine, asparagine, and glutamic acid residues; (iv) oxidation and double oxidation of methionine and proline residues; (v) phosphorylation of serine, threonine, and tyrosine residues; (vi) C-terminal amidation of lysine and arginine residues; (vii) formation of propionamide with a cysteine residue. Retention coefficient estimation was based on data from 25 mass spectrometry experiments for which identification was performed from the raw data deposited in the ProteomeXchange database. The RTP program and web service are freely available at http://lpcit.ibmc.msk.ru/RTP.