The idea that blood can serve as a coherent and interpretable system-level representation of human health has circulated for decades. What distinguishes the present moment is not conceptual novelty but infrastructural readiness. Advances in affinity proteomics, population-scale cohort design, and reproducible computational pipelines have made it possible to treat the circulating proteome not as a noisy byproduct of physiology but as a structured informational layer. María Bueno Álvez’s doctoral thesis, The Blood Proteome as a Window into Human Health and Disease, crystallizes this transition. It is not merely a contribution to biomarker discovery. It is an argument, largely implicit yet methodologically consistent, for the blood proteome as a foundational diagnostic substrate in precision medicine.
This work emerges at the intersection of three global developments. First, the maturation of ultra-high-throughput affinity-based proteomics platforms capable of measuring thousands of proteins at population scale. Second, the consolidation of biobanks and longitudinal cohorts as strategic national and transnational assets. Third, the increasing recognition that disease heterogeneity, comorbidity, and early-stage pathophysiology cannot be adequately captured through narrow case-control paradigms. The thesis responds to these developments not by proposing a single breakthrough biomarker but by assembling an integrated framework in which blood-based protein signatures are contextualized across diseases, platforms, and biological timescales.
Blood is treated throughout the thesis not as a surrogate tissue but as a systemic communication medium. Proteins in circulation encode signals derived from active secretion, immune activation, tissue remodelling, injury, and developmental processes. This heterogeneity is often presented as a challenge. Here it is operationalized as an advantage. The circulating proteome becomes readable when it is studied at scale, under controlled pre-analytical conditions, and across a sufficiently diverse phenotypic landscape. The thesis therefore departs deliberately from the classical biomarker paradigm, which assumes that specificity emerges from isolation. Instead, specificity is shown to emerge from comparison.
At the technical level, the work relies heavily on affinity proteomics, particularly Olink’s Proximity Extension Assay platform and SomaLogic’s SomaScan platform. These technologies are treated not as competing solutions but as partially overlapping measurement systems with distinct detection limits, epitope biases, and error structures. The comparative evaluation of these platforms is not ancillary. It is central to the argument that the future of blood-based diagnostics will be pluralistic, integrative, and platform-aware. Agreement between platforms is shown to be highest for disease-associated signals, suggesting that biological perturbation itself stabilizes measurement across technologies. This observation has important implications for regulatory validation, cross-cohort harmonization, and long-term data integration initiatives.
The empirical core of the thesis consists of five papers, collectively spanning early cancer detection, pan-cancer discrimination, life-course proteomics, cross-disease atlas construction, and computational infrastructure development. Together, these elements articulate a shift from biomarker hunting to proteomic cartography.
The cancer studies illustrate this shift with particular clarity. In patients presenting with non-specific symptoms, the classical diagnostic problem is not sensitivity but triage. Symptoms such as fatigue, weight loss, or malaise are common to malignant and non-malignant conditions. By comparing cancer patients not to healthy controls but to symptomatic individuals without cancer, the study reframes early detection as a probabilistic risk stratification problem. The resulting multi-protein model, built from affinity-based plasma measurements, demonstrates robust performance across independent cohorts. Importantly, the discriminative power does not rely on tumour-specific antigens alone but on integrated signals related to proliferation, immune modulation, and tissue stress.
The pan-cancer study extends this logic further. Twelve cancer types are compared against one another, rather than against a normative healthy baseline. This design reveals both shared cancer-associated protein signatures and cancer-type-specific patterns. It also demonstrates that relatively compact protein panels can reproduce the classification performance of much larger feature sets. From an industry and policy perspective, this finding is consequential. It suggests that scalable, clinically deployable assays need not sacrifice informational richness, provided they are derived from appropriately structured discovery frameworks.
Beyond oncology, the thesis constructs a human pan-disease blood atlas that spans development, adulthood, aging, and more than fifty disease states. This atlas demonstrates that while individual plasma proteomes are remarkably stable over time, systematic deviations associated with age, sex, and disease are both detectable and interpretable. During childhood and adolescence, proteomic variation is dominated by developmental trajectories rather than individual identity. In adulthood, a personal proteomic fingerprint emerges, shaped strongly by genetic factors. In disease, this fingerprint is overlaid by condition-specific perturbations that are sufficiently strong to enable multi-disease classification based on blood alone.
This approach challenges deeply rooted assumptions in clinical diagnostics. The notion of a single healthy reference range appears increasingly inadequate in a world where individual baselines are stable, disease signals are multivariate, and comorbidities are the norm rather than the exception. The thesis implicitly argues for a recalibration of diagnostic norms, one in which longitudinal self-comparison, context-aware modelling, and pan-disease benchmarking become standard.
Crucially, the work does not ignore the translational bottlenecks that have historically limited proteomic biomarkers. The reproducibility crisis in biomarker research is addressed directly through study design, validation strategy, and computational rigor. Independent cohorts, symptomatic controls, cross-platform comparisons, and open-access pipelines are repeatedly emphasized. The development of the HDAnalyzeR software package is not a methodological afterthought but a governance intervention. By standardizing exploratory and machine learning workflows, the package lowers the barrier to reproducible analysis and facilitates cross-study comparison. In a field where analytical flexibility has often undermined robustness, this kind of tooling functions as institutional infrastructure.
The integration of the resulting datasets into the Human Protein Atlas further reinforces the infrastructural character of the work. Rather than treating data as project-specific output, the thesis positions them as components of a continuously evolving public resource. The Human Disease Blood Atlas becomes not just a summary of findings but a benchmarking environment in which future biomarkers can be evaluated against a broad disease landscape. This model aligns with emerging global norms in biomedical research, where value is increasingly derived from interoperability, reuse, and longitudinal enrichment.
From a policy perspective, the thesis speaks most forcefully to the challenge of early, minimally invasive diagnosis in complex populations. Health systems worldwide are under pressure from aging demographics, multimorbidity, and constrained resources. Blood-based proteomics offers a plausible pathway toward risk stratification, screening support, and disease monitoring at scale. However, the work also makes clear that technical maturity alone is insufficient. Clinical adoption will require alignment with regulatory frameworks, reimbursement models, and laboratory infrastructure. Affinity-based proteomics platforms must transition from exploratory tools to standardized diagnostic instruments, a shift that will demand collaboration between technology vendors, clinical laboratories, regulators, and health systems.
The global dimension of this challenge is implicit but unmistakable. Large-scale plasma proteomics efforts are underway in Europe, the United States, China, and Iceland, often linked to national biobanks and public-private consortia. Companies such as Olink Proteomics and SomaLogic operate at the boundary between research and diagnostics, while pharmaceutical partners increasingly view proteomic data as both biomarker substrate and drug target validation layer. Standardization bodies and data infrastructures, including ELIXIR and international proteomics consortia, are becoming as critical as assay chemistry itself. In this context, the thesis can be read as a case study in how publicly anchored, academically driven research can shape global diagnostic trajectories.
At a more conceptual level, the work reframes the blood proteome as a living map rather than a static inventory. Proteins are not merely measured; they are situated within biological time, disease space, and technological constraint. The emphasis on pan-disease comparison is especially important. It reveals that many proteins traditionally labelled as disease-specific are, in fact, markers of shared pathophysiological processes. Specificity emerges not from exclusivity but from patterning. This insight has far-reaching implications for how diagnostic tests are designed, validated, and interpreted.
The thesis remains largely internal to this constructive vision. It does not dwell on unresolved regulatory debates, reimbursement hurdles, or ethical tensions beyond data privacy and re-identification risk. Yet these issues are not absent. They are embedded in the architecture of the work. By demonstrating that plasma proteomic profiles can act as personal molecular fingerprints, the thesis underscores the sensitivity of such data. By releasing datasets through controlled public platforms, it implicitly engages with questions of consent, governance, and responsible reuse.
What ultimately distinguishes this body of work is its coherence. Each component, from cancer diagnostics to software development, reinforces a central thesis: that the circulating proteome, when studied systematically and at scale, constitutes a powerful and actionable layer of biomedical information. The promise is not a single test or platform, but an ecosystem in which blood-based proteomics supports earlier detection, finer stratification, and more adaptive healthcare systems.
For industry leaders, the message is that value will accrue to those who invest in robustness, interoperability, and longitudinal data integration rather than isolated assays. For policymakers, the implication is that blood-based proteomics should be treated as strategic diagnostic infrastructure, deserving of coordinated investment and regulatory foresight. For the scientific community, the thesis offers a model of how methodological rigor, open resources, and biological ambition can be aligned.
The blood proteome is no longer merely a window. It is becoming a framework through which health and disease can be measured, compared, and ultimately managed at population scale.
References
Álvez, M. B., Bergström, S., Kenrick, J., et al. (2025). A human pan-disease blood atlas of the circulating proteome. Science, 390(6779), eadx2678.
Álvez, M. B., Edfors, F., Von Feilitzen, K., et al. (2023). Next generation pan-cancer blood proteome profiling using proximity extension assay. Nature Communications, 14(1), 4308.
Wannberg, F., Álvez, M. B., Qvick, A., et al. (2025). Plasma protein profiling predicts cancer in patients with nonspecific symptoms. Nature Communications, 17(1), 151.
Antonopoulos, K., Johansson, E., Kenrick, J., et al. (2026). HDAnalyzeR: Streamlining data analysis for biomarker research. Bioinformatics Advances, 6(1), vbag020.
Uhlén, M., Fagerberg, L., Hallström, B. M., et al. (2015). Tissue-based map of the human proteome. Science, 347(6220), 1260419.
Wik, L., Nordberg, N., Broberg, J., et al. (2021). Proximity Extension Assay in combination with next-generation sequencing for high-throughput proteome-wide analysis. Molecular & Cellular Proteomics, 20, 100168.
Bueno Álvez, M. (2026). The blood proteome as a window into human health and disease. Doctoral thesis, KTH Royal Institute of Technology.
