Investigation and Highly Accurate Prediction of Missed Tryptic Cleavages by Deep Learning

Comparison of N-Glycopeptide to Released N-Glycan Abundances and the Influence of Glycopeptide Mass and Charge States on N-Linked Glycosylation of IgG Antibodies

We report the comparison of mass-spectral-based abundances of tryptic glycopeptides to fluorescence abundances of released labeled glycans and the effects of mass and charge state and in-source fragmentation on glycopeptide abundances. The primary glycoforms derived from Rituximab, NISTmAb, Evolocumab, and Infliximab were high-mannose and biantennary complex galactosylated and fucosylated N-glycans. Except for Evolocumab, in-source ions derived from the loss of HexNAc or HexNAc-Hex sugars are prominent for other therapeutic IgGs. After excluding in-source fragmentation of glycopeptide ions from the results, a linear correlation was observed between fluorescently labeled N-glycan and glycopeptide abundances over a dynamic range of 500. Different charge states of human IgG-derived glycopeptides containing a wider variety of abundant attached glycans were also investigated to examine the effects of the charge state on ion abundances. These revealed a linear dependence of glycopeptide abundance on the mass of the glycan with higher charge states favoring higher-mass glycans. Findings indicate that the mass spectrometry-based bottom-up approach can provide results as accurate as those of glycan release studies while revealing the origin of each attached glycan. These site-specific relative abundances are conveniently displayed and compared using previously described glycopeptide abundance distribution spectra "GADS" representations. Mass spectrometry data are available from the MAssIVE repository (MSV000093562).

Large-Scale Profiling of Unexpected Tryptic Cleaved Sites at Ubiquitinated Lysines

Trypsin specifically cleaves the C-terminus of lysine and arginine residues but often fails to cleave modified lysines, such as ubiquitination, therefore resulting in the uncleaved K-ε-GG peptides. Therefore, the cleaved ubiquitinated peptide identification was often regarded as false positives and discarded. Interestingly, unexpected cleavage at the K48-linked ubiquitin chain has been reported, suggesting the latent ability of trypsin to cleave ubiquitinated lysine residues. However, it remains unclear whether other trypsin-cleavable ubiquitinated sites are present. In this study, we verified the ability of trypsin in cleaving K6 and K63 besides K48 chains. The uncleaved K-ε-GG peptide was quickly and efficiently generated during trypsin digestion, whereas cleaved ones were produced with much lower efficiency. Then, the K-ε-GG antibody was proved to efficiently enrich the cleaved K-ε-GG peptides and several published large-scale ubiquitylation datasets were re-analyzed to interrogate the cleaved sequence features. In total, more than 2400 cleaved ubiquitinated peptides were identified in the K-ε-GG and UbiSite antibody-based datasets. The frequency of lysine upstream of the cleaved modified K was significantly enriched. The kinetic activity of trypsin in cleaving ubiquitinated peptides was further elucidated. We suggest that the cleaved K-ε-GG sites with high post-translational modification probability (≥0.75) should be considered as true positives in future ubiquitome analyses.

Toward an Integrated Machine Learning Model of a Proteomics Experiment

In recent years machine learning has made extensive progress in modeling many aspects of mass spectrometry data. We brought together proteomics data generators, repository managers, and machine learning experts in a workshop with the goals to evaluate and explore machine learning applications for realistic modeling of data from multidimensional mass spectrometry-based proteomics analysis of any sample or organism. Following this sample-to-data roadmap helped identify knowledge gaps and define needs. Being able to generate bespoke and realistic synthetic data has legitimate and important uses in system suitability, method development, and algorithm benchmarking, while also posing critical ethical questions. The interdisciplinary nature of the workshop informed discussions of what is currently possible and future opportunities and challenges. In the following perspective we summarize these discussions in the hope of conveying our excitement about the potential of machine learning in proteomics and to inspire future research.

DNA and protein analyses of hair in forensic genetics

Hair is one of the most common pieces of biological evidence found at a crime scene and plays an essential role in forensic investigation. Hairs, especially non-follicular hairs, are usually found at various crime scenes, either by natural shedding or by forcible shedding. However, the genetic material in hairs is usually highly degraded, which makes forensic analysis difficult. As a result, the value of hair has not been fully exploited in forensic investigations and trials. In recent years, with advances in molecular biology, forensic analysis of hair has achieved remarkable strides and provided crucial clues in numerous cases. This article reviews recent developments in DNA and protein analysis of hair and attempts to provide a comprehensive solution to improve forensic hair analysis.

Improving SWATH-MS analysis by deep-learning

Data-independent acquisition (DIA) of tandem mass spectrometry spectra has emerged as a promising technology to improve coverage and quantification of proteins in complex mixtures. The success of DIA experiments is dependent on the quality of spectral libraries used for data base searching. Frequently, these libraries need to be generated by labor and time intensive data dependent acquisition (DDA) experiments. Recently, several algorithms have been published that allow the generation of theoretical libraries by an efficient prediction of retention time and intensity of the fragment ions. Sequential windowed acquisition of all theoretical fragment ion spectra mass spectrometry (SWATH-MS) is a DIA method that can be applied at an unprecedented speed, but the fragmentation spectra suffer from a lower quality than data acquired on Orbitrap instruments. To reliably generate theoretical libraries that can be used in SWATH experiments, we developed deep-learning for SWATH analysis (dpSWATH), to improve the sensitivity and specificity of data generated by Q-TOF mass spectrometers. The theoretical library built by dpSWATH allowed us to increase the identification rate of proteins compared to traditional or library-free methods. Based on our analysis we conclude that dpSWATH is a superior prediction framework for SWATH-MS measurements than other algorithms based on Orbitrap data.

Improved Prediction Model of Protein Lysine Crotonylation Sites Using Bidirectional Recurrent Neural Networks

Histone lysine crotonylation (Kcr) is a post-translational modification of histone proteins that is involved in the regulation of gene transcription, acute and chronic kidney injury, spermatogenesis, depression, cancer, and so forth. The identification of Kcr sites in proteins is important for characterizing and regulating primary biological mechanisms. The use of computational approaches such as machine learning and deep learning algorithms have emerged in recent years as the traditional wet-lab experiments are time-consuming and costly. We propose as part of this study a deep learning model based on a recurrent neural network (RNN) termed as Sohoko-Kcr for the prediction of Kcr sites. Through the embedded encoding of the peptide sequences, we investigate the efficiency of RNN-based models such as long short-term memory (LSTM), bidirectional LSTM (BiLSTM), and bidirectional gated recurrent unit (BiGRU) networks using cross-validation and independent tests. We also established the comparison between Sohoko-Kcr and other published tools to verify the efficiency of our model based on 3-fold, 5-fold, and 10-fold cross-validations using independent set tests. The results then show that the BiGRU model has consistently displayed outstanding performance and computational efficiency. Based on the proposed model, a webserver called Sohoko-Kcr was deployed for free use and is accessible at https://sohoko-research-9uu23.ondigitalocean.app.