Recent advances, applications and challenges in proteomics

Proteomics is one of the most rapidly developing hot fields in biopharmaceuticals.
The study of proteomics can not only deepen the understanding of molecular processes that support biological states across cells, tissues and whole organisms, but also help humans, animals and plants.
Various fields of scientific research, including biology, personalized medicine, and forensic science, are developing rapidly
.
These R&D gains are largely due to advances in proteomics technology, data processing capabilities, and data sharing
.
In this article, we join leading proteomics experts from around the world to explore the latest advances in proteomics and their broader potential impact
.
I.
Development of key technologies in proteomics 1.
Mass spectrometry proteomics The analytical techniques used in proteomics research can be roughly divided into two categories: low-throughput and high-throughput
.
For decades, mass spectrometry (MS) has been the most widely used "gold standard" in high-throughput analysis
.
Historically, a key issue for mass spectrometry-based proteomics has been sensitivity and specificity, but in recent years the landscape of mass spectrometry has changed dramatically
.
Scientists can now also delve deeper into proteomics with the introduction of mass spectrometry instruments with high speed, sensitivity and specificity from suppliers
.
Asked about the breakthroughs in mass spectrometry proteomics, Prof.
Matthias Mann, Research Director and Group Leader of the Proteomics Program at the Novo Nordisk Foundation Center for Protein Research, highlighted Aebersold's laboratory rate in Data Independent Acquisition (DIA).
progress
.
Unlike its sister technique Data Correlation Analysis (DDA), DIA dissociates in the second cycle (MS2) all precursor ions produced in the first cycle of tandem MS (MS1), which also provides researchers with more Good unbiased analysis capability, greater proteome coverage and higher reproducibility
.
In recent years, the use of DIA-based mass spectrometry has also continued to grow in proteomic research, especially in oncology.

.
In addition, DIA technology has also made positive progress in the field of neuroscience proteomics, including the discovery of information related to Alzheimer's disease
.
2.
Aptamer-based proteomics Although mass spectrometry has dominated the field of proteomics research for many years, the recent emergence of “second-generation” proteomics platforms utilizes novel aptamer-based technologies
.
Dr.
Benjamin Orsburn, a researcher at the Johns Hopkins University School of Medicine, also said that although LC-MS has had a monopoly on proteomics for decades, this is clearly no longer the case
.
Aptamers are short single-stranded (ss) DNA molecules capable of unique match confirmation that selectively bind to biological targets
.
In areas such as biomarker discovery, the technology has advantages such as specificity and selectivity, and thus outperforms MS proteomics, which is limited by its dynamic range
.
Previously, the researchers studied aptamer-based proteomics in 1895 women from the Framingham Heart Study to identify biomarkers of cardiac remodeling and heart failure events
.
The trial results showed that 17 proteins were found to be associated with echocardiographic features and 6 proteins were associated with heart failure events, and further analysis using genetic variation data further supported these findings
.
Orsburn said the use of aptamer technology will be less affected by the absolute protein copy number in the cell than with LC-MS technology
.
However, until a higher percentage of the proteome can be identified, mass spectrometry proteomics will likely remain the method of choice for now, with aptamer-based techniques often being complementary and complementary
.
Recently, the ideal protein sequencing platform proposed by researchers uses barcoded DNA aptamers to identify the terminal amino acids of peptides, and linking to next-generation sequencing chips may be a compromise
.
Orsburn said the field's full R&D potential still takes time to realize
.
2.
Artificial intelligence and proteomics 1.
Artificial intelligence and drug discovery proteomics The application of artificial intelligence (AI) in proteomics has profoundly changed the research and development work in the field of drug discovery
.
For researchers, gaining a deep understanding of how and why specific proteins can interact is critical to advancing cell biology, developing new drugs, and identifying the mechanisms by which drugs trigger treatments and side effects, but it is by no means easy
.
Octavian-Eugen Ganea, a postdoctoral researcher at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL), said that in order to understand how interacting proteins are connected to each other, all possible combinations of connections must be tried, either manually or by computer, in order to find the most plausible one.
, which would be a very time-consuming process if done manually without the aid of artificial intelligence
.
A variety of commercial protein docking methods are available, but they all rely on candidate sampling, templates, and task-specific characteristics of precomputed grids, all of which add additional computational time
.
The MIT Ganea team recently published EquiDock, a new deep learning model that takes the 3D structures of two proteins and can directly identify which regions are likely to interact
.
The EquiDock model is able to capture complex docking patterns from about 41,000 protein structures, using a geometrically constrained model with thousands of parameters that can be dynamically and automatically adjusted during computation
.
Once trained, the EquiDock model was cross-compared with four other existing docking software and was able to predict the final protein complex in one to five seconds, 80 to 500 times faster than existing software
.
Ganea said that fast computational scanning of drug side effects is very necessary, which can significantly reduce the search scope, otherwise, even if the global manual testing resources are integrated, satisfactory results cannot be obtained.

.
He emphasized that combining the EquiDock model with other protein structure prediction models is expected to further aid its applications in drug design, protein engineering, antibody generation, and mechanism of action research
.
2.
Artificial intelligence and mass spectrometry proteomics AI-based approaches will also help researchers gain more R&D insights from the data obtained
.
Mass spectrometry experiments require database searches or spectral library matching to identify specific proteins
.
Not only is the entire process particularly time-consuming, but some proteins may also be misidentified or missed, which seem even more unavoidable for DIA mass spectrometry, which relies on the generation of spectral libraries through DDA analysis
.
To this end, researchers have now established a variety of deep learning methods capable of predicting spectral and peptide properties, including but not limited to Prosit, DeepMass, and DeepDIA, which are expected to optimize the predicted spectral library of DIA methods and bring the field of proteomics closer to develop in a better direction
.
3.
Artificial intelligence and non-mass spectrometry proteomics AI can also assist in the development of non-mass spectrometry proteomics, a field that is invaluable for understanding pathologies characterized by tangled, aggregated proteins, such as Alzheimer's disease.
or missing
.
Key methods employed in this field include microscopy and Förster resonance energy transfer (FRET), and the R&D process requires significant time and sufficient expertise to interpret large datasets
.
To overcome this data conundrum, researchers at the Novo Nordisk Foundation Center for Protein Research, led by Professor Nikos Hatzakis, recently created the DeepFRET model
.
DeepFRET is an AI-powered machine learning algorithm that can automatically identify protein motion patterns, classifying datasets in seconds (typically days if done entirely manually)
.
The future development of AI in proteomics also requires that AI platforms must comply with relevant standards, data reporting and sharing aspects to achieve cross-group synchronization
.
The recently published Data, Optimization, Models, Evaluation (DOME) recommendations for conducting and reporting machine learning in proteomics and metabolomics may help reshape the future direction of the field
.
3.
Forensic Medicine and Proteomics The wider application of proteomics also benefits from the technological advances discussed earlier, such as the "DNA revolution" in the late 20th century that greatly contributed to the development of forensic science, and now proteomics seems promising to produce similar effects
.
In this regard, Glendon Parker, Ph.
D.
, the inventor of protein-based human identification technology, said that in general, the current impact of proteomics on forensics is limited due to technical, legal, financial and cultural factors, however, The adoption and incorporation of new methods in criminal investigations and prosecutions will be a fundamental driver
.
Parker added that proteomics has unique inherent advantages, such as proteins being more stable than DNA and being able to contain specific identifying information like DNA
.
In cases where DNA nucleic acids are degraded, proteomics can be used to identify body fluids, gender, ethnicity, and to estimate approximate time of death using muscle, bone, and lysate samples
.
Parker emphasizes that while the true implementation of proteomics in forensics has always been challenging, in the future proteomics has the potential to significantly change the way forensic evidence is processed and analyzed
.
In the short term, the field could assist DNA technology in forensic areas where DNA is difficult to provide clear evidence
.
4.
Challenges and future prospects of proteomics 1.
Urgent need for open sharing and strengthening of global cooperation At present, the biggest limitation facing the field of proteomics is the complexity of technology, which requires R&D personnel to operate proficiently in the lengthy workflow of proteomics Sophisticated technology and specialized software
.
Although proteomics has made rapid progress in terms of sensitivity and speed, and research results have also made corresponding progress, there is a huge price behind it.

.
Rigorous, deep-coverage mass spectrometry experiments, especially those on complex biological samples, require a significant amount of mass spectrometry time, so developers must constantly juggle R&D cost, coverage, and sample size, Johnston said.
Weigh the pros and cons
.
This is also one of the current difficulties that limit the wider application of proteomics.
Parker also emphasized that these difficulties limit technological innovation, resulting in a large number of promising emerging technologies (including proteomics), which are ultimately underutilized
.
In the past decade, there has been a growing call for openness, sharing, and collaboration in the field of proteomics
.
Concrete initiatives to increase accessibility and sustainability are already emerging, such as the European Proteomics Infrastructure Consortium (EPIC-XS), which brings together some of Europe's leading laboratories and scientists to Trial data are pooled and shared
.
The shared resources are also not limited to mass spectrometry-based proteomics, the cellular analysis facility at the KTH access site also opens up expertise in antibody-based imaging
.
2.
Difficulties in advancing clinical applications There are still a number of key challenges to overcome before proteomics can be formally identified as a clinical pillar, and these challenges depend on the specific application in clinical proteomics
.
Mass spectrometry-based proteomics needs to become more powerful and easy to use if it is to be used at scale in the clinic, Mann said, and while many R&D groups have attempted to achieve this by moving to high-flow chromatography systems, current results have not Not ideal as this will cause sensitivity to suffer
.
At present, although analytical technologies have improved the ability to dig deep into the proteome to a certain extent, the amount of data generated is also growing simultaneously, which also brings additional bottlenecks to the clinical application of proteomics, including how to process large amounts of data.
, and how to formulate biological and clinical hypotheses on the basis of large-scale datasets
.
In addition, for a comprehensive understanding of human health and disease, proteomic data must often be used in conjunction with other omics counterparts, such as metabolomics, genomics, and transcriptomics
.
In addition, ethical issues are also seen as one of the challenges for proteomics to enter the clinic
.
Proteomic analysis can provide information beyond the original diagnosis, and there are still regulatory gaps in how clinicians will process this data
.
While some lessons can be learned from previous clinical implementations of clinical genomics, the field of proteomics is an entirely different situation, making it difficult to directly apply previous standards when developing regulatory frameworks and guidelines
.
However, despite the current limitations, leading experts are confident in the future of the field, and Mann pointed out that mass spectrometry detection technology will continue to develop in clinical applications due to its inherent specificity
.