One of my volunteer jobs in HUPO is to interview scholars and broadcast their thinkings to the community. This is a recorded 1v1 interview with Dr Ruedi Aebersold. The content was originically released by HUPO official Wechat account: HUPO-Wechat. Permission to reproduce the article was granted.
Original Title: HUPO ReCONNECT 2021 interview: Ruedi Aeberold– Xiao Liang
Interviewer & Interviewee
To offer ECRs a better opportunity to have deep communication with established experts in the field of proteomics, HUPO WeChat Official Account has launched a series of interviews conducted by motivated ECRs from China with relevant educational backgrounds. In each interview, HUPO ReConnect 2021 speakers will discuss issues of concern with ECRs, ranging from academic challenges, career planning to industry prospects.
For the final round of our interview project, Xiao Liang, a Ph.D. student from Westlake University, interviewed Prof. Ruedi Aebersold, one of the pioneers in the field of systems proteomics. Prof. Aebersold is known for having developed a series of methods that have been widely used in analytical protein chemistry and proteomics. He serves on the Scientific Advisory Committees of numerous academic and private-sector research organizations, as well as being a member of several editorial boards in the fields of protein science, genomics, and proteomics.
Prof. Aebersold shared with Xiao and ECRs his advice on ECRs, opinions on MS-free proteomic technologies, and the topic in the HUPO ReConnect 2021 meeting.
Q1 Brief introduction of your recent interest in proteomics.
It is important to move proteomics out of a purely technical domain into a biological or clinical discipline. We have many discussions of very impressive technical advances at conferences, e.g., at HUPO. Scientists achieve better sensitivity and higher resolution and throughput to analyze higher numbers of samples. Those advances are about identifying more proteins per sample. This is a very important goal of proteomics, but by itself not sufficient.
Because proteins are the most important biologically functional molecules, we have been working on ways to translate proteomic measurements into biological knowledge. As one approach in that direction, we have been very interested in determining how the proteome of a cell reacts to changes. Those changes could be caused by genomic changes in diseases, like cancer cells acquiring mutations, or by external changes, like drugs, mechanical or chemical stress. We are interested to know what are the mechanisms by which the cell detects these changes and how it reacts to them. This question is very important for molecular medicine, for instance in cancer biology where you can detect numerous genomic mutations that are present in a tumor of a particular patient. But how do these mutations generate a disease phenotype? What type of disease? Is the disease resistant to therapy or not? We still don’t know the answers to these questions in a general way, i.e., we do not have a model that predicts how a cell reacts to a particular perturbation. We are interested in determining principles of how genomic variation and other perturbations affect phenotypes, by determining the state of the proteome in cells at different states.
Q2 What’s your topic in the HUPO ReConnect 2021 meeting?
I am preparing a presentation for the HUPO ReConnect 2021 meeting where I will discuss the richness of proteomic data in terms of biological information. When you look through the literature in proteomics or when you go to a conference like HUPO, most proteomics studies focus on identifying proteins in one or many samples and determining their abundance. This is usually done by bottom-up proteomics. When you take a step back and ask, what we really do in this context is to identify and quantify one or a few out of potentially many proteoforms that can be generated from each of the approximately 20,000 protein-coding genes. This has been already useful in basic science, as well as in clinical research. But I think we must realize that proteins contain a huge amount of additional information beyond their sequence and abundance. For instance, proteins can be modified and can also be processed e.g., by N-terminal truncation, they can change their structure and they can associate with one or many macromolecular complexes. Many of those properties or changes thereof that proteins can experience are very biologically relevant.
So, I plan to talk about how we can expand proteomics to measure these other’ properties of proteins that provide additional functional information more systematically and how we can relate these properties to phenotypes or functions.
Q3 Any advice for ECRs?
– For ECRs Who are Ph.D. students and postdocs:
For Ph.D. students and postdocs alike, you need to understand that the world of science is very competitive. There are many intelligent and dedicated, hardworking young scientists. If a postdoc or Ph.D. student’s goal is to become an independent scientist, they need to undertake certain strategic steps and maximize the chance that they will succeed in eventually starting their own lab.
Those certain strategic steps can be broken down into two or three categories.
For the first, you need to demonstrate that as a young scientist you can create meaningful research results. You should avoid selecting projects of a type that has been done ten times and do it for the 11th time with very minimal alterations. The results from this kind of “me too” science won’t get you into a strong position.
In contrast, you can also not afford to take on a project which may have a long lead of time of e.g., 10 years. You will not have a chance in your Ph.D./PD period to complete such a project.
So, the project you have chosen should be ambitious, not overly ambitious, feasible, but not trivial. If the project works, it can create results that others may find interesting and moves along the field of research which means that you will be able to publish the results in a relatively visible or highly visible journal.
The choice of project is very difficult for a Ph.D. student, and a bit less difficult for a postdoc who already has gained some experience. In either case, suggestions and guidance from the advisor are critically important. The advisor or advisor group should be able to guide postdocs to meaningful research projects with a relatively high chance of success. For beginning Ph.D. students, planning a research project is very hard to do because they lack experience. In that regard, the group of advisors is very important in the process of defining a research project.
For Ph.D. students who have an academic ambition and seek out a lab for postdoctoral research, I advise them to look at the track record of the group they’d like to join to see how many incoming postdocs have succeeded in starting their own laboratory from that group. If no one from that group managed to go this path for the last e.g., 10 or 15 years, then the chance for the incoming postdoc to successfully start his/her own lab is relatively small. On the other hand, if success happens in that group occasionally, your chances of making the big step are intact. This means the scientific environment is conducive to a decisive career step, of course apart from the skills of the individual researcher.
Another advice for Ph.D. students is to talk to other people in the field as often as possible. I think HUPO is a great opportunity to talk to scientists who are experienced in proteomics. This allows you to collect as much information and points of view as possible. Sometimes, a good discussion of half an hour can save you several months of time in the lab.
– For the independent group leaders:
Independent group leaders have already managed to convince a review or appointment committee that they are the best person to be appointed for a position. Starting group leaders usually have a certain amount of time to reach the next stage, typically five years to seven years. This seems like a long time, but in reality, it is a very short time to establish one’s own research program and to show initially published accomplishments. My advice to starting group leaders is to be very careful with time management. E.g., it may take a year to get students recruited and trained and it may take another year to get finished research results through the review process and published. I think it is a good strategy to start as an independent scientist with some moderately ambitious projects with a relatively high probability of success and a high probability of publishing initial papers after three or four years. Such projects also establish the skills of a young group leader to guide a student or a postdoc through a project towards publication.
On top of that, group leaders need to formulate a bigger vision and more ambitious goals and projects. If such a project succeeds, it will generate highly significant results required to advance the career. The challenge is to find the right balance between getting traction as a group and achieving an ambitious goal. At last, I would suggest to seek advice, talk to more experienced colleagues, and seek their input. They generally are very glad to share the experience from their own career.
Q4 What’s your opinion on MS-free proteomic technologies, and do you think they could replace MS in the future?
Scientists in the field, especially young investigators will ask themselves, whether they are doing work that will be irrelevant in e.g. 5 or 10 years? If the answer is yes, then they must think about how to deal with that. I think it is actually a very important question for ECRs in general but particularly for those working in technologically fast-moving fields.
Personally, I don’t think that mass spectrometry will be overtaken by other techniques anytime soon. This is not to say that mass spec-free techniques are not very powerful. On the contrary, they are rapidly advancing, particularly in the commercial arena. And those techniques, particularly affinity-reagent-based techniques have a substantial advantage over mass spectrometry. Once you have suitable reagents, like an antibody or an aptamer or combinations of antibodies, these reagents can be validated and then used for very large numbers of measurements. These methods have therefore the great advantage that they are scalable. The readout is relatively simple, and the reagents are relatively cheap to use. You do not need a million-dollar instrument to do perform affinity-reagent-based assays. To some extent, mass spec-free techniques are very already well established in laboratories and in commercial enterprises. Generally, these methods measure protein abundance and the presence of a protein in a sample, like bottom-up proteomics. In this domain, affinity-based methods and MS methods directly compete and they have different performance characteristics in terms of throughput, cost, precision, and error management.
As we discussed above, there’s much more to proteomics than detecting and quantifying proteins. For instance, how does the protein change its fold when it is phosphorylated, or how does a protein change its association with a complex if an intron is spliced out are examples of effects with clear functional consequences. E.g., if a kinase is phosphorylated in the activation loop, access to the ATP binding site may be obscured or closed causing the kinase to be no longer active. For non-mass spec methods, it will be challenging and potentially very expensive to develop reagents that can systematically determine such functionally relevant events because one would have to generate an antibody that detects a certain shape of a protein, or an antibody or affinity reagent that will determine whether two proteins are interacting. The mass spectrometric measurement of such events is much faster and straightforward to achieve. Therefore, I think mass spectrometry will remain the method of choice for the exploration of the richness of biological information of the proteome due to its unique strength, versatility, and flexibility of its applications.
Q5 Genomics is evolving towards single-cell spatial research. Do you think we are ready for spatial proteomics when the resolution and throughput are limited?
Single-cell proteomics is still in an early phase and is technically very challenging. For its further development, it may be interesting to look at the development of single-cell transcriptomics. The transcript sequencing on single cells initially did not really cover the complete transcriptome. Over time, since the inception of the method, the coverage achieved, as well as the number of cells that can be processed have increased a lot. In proteomics, it is presently possible to measure a certain number of proteins (in the range of 1000-2000) from a single cell and to do this with relatively low throughput. It can be expected that also in proteomics the number of analytes measured per cell and the number of cells analyzed per study will increase. Whether and how single-cell proteomics can be scaled to a very large number of cells which can be analyzed by transcriptomics is not so clear at present. So, I think there is still a certain number of technological developments required, and this will probably take some time.
But the issue should not just be discussed from a technical side but also from the point of view of biological objectives. Generally, single-cell transcriptomic or single-cell genomics have been very powerful in determining lineages of cells in a tissue. For instance, in tumors where a sequence of mutations occurred in cell lineages, single-cell data can be used to build a dependency tree of tumor cells to determine how cells evolve in the environment of the tumor. This has been done perfectly well with transcripts data, and I don’t think that protein data would really add to that. Single-cell proteomics, therefore, should not just be a conceptual carbon copy of single-cell transcriptomics, but rather seek its own questions where transcriptomes cannot reach. All this will take a while. I am excited about these developments, but it will require some time and effort to develop robust technologies and then to formulate suitable research questions.
Q6 You’ve recently joined the Chinese Human Proteomic Project: Proteomics-Driven Precision Medicine (PDPM) as an advisory board. What’s your expectation and concern on this huge project?
As far as I know, the project is still in the phase where it is being conceptualized, and many steps need to be taken to make the project a reality. There are many very ambitious plans and ideas but they will need to be reduced to practice and prioritized. The project has a very high potential for success if the available resources can be deployed in a structured manner on ambitious project goals. Simply by the scale of the project and the energy of the project team, this will be an exciting project. I don’t know any other project of this scale or ambition anywhere else. I am very excited to be part of the team that helps to shape the ideas. I think the focus on precision medicine is highly timely, and the expected results have the potential to accomplish a lot not only in terms of basic science and technology advancements but also for the patient.
Usually, with such large-scale projects, the concerns and the threats include the following: How are the goals defined? Who is on board? How are the project and governance structures being established? How can a large number of scientists be motivated to work towards these goals? Who is leading the project and how can the expectations of the financial supports be managed? How is credit from accomplishments distributed? Therefore, challenges for the project include the formulation of common goals and getting the team built, motivated and agreeable to pursue this common goal. In precision medicine, the overall goal is obviously quite well defined, to associate the patient with the optimal treatment based on molecular constellations, but the path towards reaching the goal requires a lot of thought.
The second concern is how a balance can be found in data generation and the exploitation of data for clinically relevant insights. Many participating groups will measure all kinds of sample sets and try to translate this information or data into useful results for patients. I have personally been involved in a smaller scale but very advanced project, the tumor profiler project in Zürich/Basel Switzerland. The most significant challenge we faced in that project was not the generation of data but rather how to process the massive amounts of information generated from each patient of a cohort into a report that is actionable by oncologists in the clinic.
The third concern is the timescale available to achieve the goals of the project. It is a highly complex project with technically and scientifically difficult questions. The scientists will do their best to reach these very ambitious goals but both funders and scientists need to recognize that it will take some time to achieve success and that success will come in incremental steps, scientifically and clinically. E.g., it is not reasonable to expect that precision medicine will be implemented in every hospital within 3-5 years.
“Thinking wisely before thinking big.” Ruedi Aebersold shared intellectual aspirations towards the community and offered concrete suggestions to ambitious ECRs at the Talking to ECRs session. Additionally, Ruedi’s talk highlighted arguable thinkings towards future proteomics. The majority of his concerns did not arise from technical capability but biomedical practicability. With the rapid development of proteomic science, these considerations shall be solved shortly. It is those who keep reflecting that will continuously raise innovative ideas and maintain the sustainable advancement of science.