How fast is fast enough in mass spectrometry? For laboratories handling large sample cohorts, the answer may seem simple: the more samples processed, the better. But the experts we spoke to at ASMS 2026 compilation argue for a more measured approach. Throughput is important, but only when it supports the scientific question without weakening data quality.
That distinction becomes significant in population-scale research and exposome studies, where scientists may need to analyze samples from hundreds of thousands of people. These projects require scalable methods. Yet pushing a workflow to its limits can force trade-offs in analytical depth or confidence. Rather than chasing the highest possible number, laboratories should ask whether a method delivers the speed and quality their application demands.
The discussion applies the same thinking to instrument purchases. Acquisition price often drives the decision, especially when budgets remain tight. However, the lowest upfront cost may not produce the best long-term return. A system that improves uptime, reduces manual work, or processes samples more efficiently could save money over its lifetime. The more useful question is not simply “What does it cost?” but “What value does it bring to the operation?”
This focus on value also shapes emerging areas of research, such as growing interest in intact proteins and proteoforms, an approach described as “long-read proteomics.” Traditional bottom-up proteomics breaks proteins into smaller peptides before analysis. Measuring whole proteins can preserve information about the different molecular forms found in the body, offering another route to understanding biological function and disease. Once viewed as a niche field, proteoform research now occupies a more prominent place at ASMS.
Artificial intelligence runs through much of the conversation. Some laboratories already use machine learning to process data across large sample sets with less manual intervention. AI tools could also support method development, generate reports, flag questionable peak integrations, and help scientists focus their attention on results that need review.
Instrument maintenance offers another practical use. By learning from previous faults, an AI system could help diagnose problems, recommend preventive action, or ensure that a service engineer arrives with the right replacement part. These applications could reduce downtime and help laboratories maintain consistent performance.
Experts stop short of treating AI as a replacement for scientific judgment. They see it as an assistant that can handle repetitive work and guide decisions while keeping a person in control. That oversight matters because every model depends on its training data. In protein science, tools such as AlphaFold have transformed structure prediction, but they have not answered every question. Protein-folding kinetics, intrinsically disordered regions, and protein interactions still present major challenges. Experimental measurements remain essential for filling those gaps and building better models.
The takeaway is one of balance. Laboratories need speed, but they also need trustworthy data. They need advanced technology, but they must justify the investment. AI can make mass spectrometry workflows more efficient, yet scientists still provide the context and judgment that turn results into insight.
