Comparative biosignatures with systemic retrievals

Tereza Constantinou, Oliver Shorttle, Miles Cranmer, Paul B. RimmerMonthly 2026 · DOI

TL;DR

This paper introduces a novel "comparative biosignature" framework that uses systemic retrievals to analyze multiple planets within a single system. By establishing an "abiotic baseline" from planets that clearly follow non-biological processes, the authors can identify anomalies in other planets that might indicate life, offering a more robust method for detecting extraterrestrial life.

What Problem Does This Paper Address?

The search for life beyond Earth, particularly through detecting biosignature gases in exoplanet atmospheres, faces a significant challenge: attribution. It's difficult to confidently identify a gas as a sign of life when non-biological (abiotic) processes can mimic or produce similar signals. This problem is amplified by the inherent complexity and stochastic nature of planetary atmospheric evolution, making it hard to establish a universal "normal" for a planet's atmosphere. Current methods often struggle to rule out abiotic explanations, especially for individual planets where context is limited. This paper aims to overcome this by proposing a multi-planet, system-wide approach to define a local "abiotic baseline" for comparison.

Methods

The core of the proposed method is systemic retrieval and the establishment of an abiotic baseline. This involves:

  • Hierarchical Bayesian Models (HBMs): These models analyze all planets within a system simultaneously, allowing shared, unobservable (latent) parameters (like initial volatile budgets or migration histories) to be constrained by the system as a whole. This helps overcome the limitations of studying individual planets in isolation.
  • Abiotic Baseline Construction: The HBM is used to model the expected atmospheric conditions for planets in a system based on known abiotic processes. Planets that deviate significantly from this baseline are flagged as potential anomalies. This baseline is iteratively refined by removing influential outliers using Pareto k diagnostics.
  • Bayesian Leave-One-Out Cross-Validation (LOO-CV): This statistical technique is used to assess how well models predict data from planets not included in the model training. This helps identify planets that are statistical outliers from the established abiotic baseline.
  • Comparative Biosignature Definition: Anomalies are then evaluated against both abiotic and biotic planetary evolution models. A "comparative biosignature" is identified when a biotic model provides a significantly better fit to the anomalous planet's data than the abiotic model (quantified by a difference in elpdLOO scores, ΔEL > 4).

Datasets

The paper does not introduce new datasets. Instead, it proposes a framework for analyzing existing and future observational data from exoplanet systems. The Solar System is used as a primary example and test case for the proposed methodology, leveraging data from planets like Earth, Venus, and Mars.

Key Insights and Ideas

  • Systemic Context is Crucial: Analyzing planets within their host system provides a more robust "local" reference point for identifying anomalies than relying on universal models or single-planet studies.
  • The "Abiotic Baseline" as a Local Norm: By defining what constitutes "normal" abiotic behavior within a specific planetary system, we can more confidently identify deviations that might be biological.
  • "Unknown Unknowns" as Scientific Opportunities: The framework explicitly accounts for anomalies that defy both current abiotic and biotic models, flagging them as "unknown unknowns" that point to gaps in our understanding or novel forms of life.
  • Comparative Habsignatures: The framework can be extended beyond biosignatures to "habsignatures" (indicators of habitability, like the presence of oceans), by defining baselines based on the absence of those specific conditions.

Key Results

The paper is primarily a theoretical and methodological proposal. It does not present quantitative results from applying the framework to exoplanet data. Instead, it outlines the methodology and provides conceptual examples, such as:

  • Earth as an Anomaly: The paper uses Earth's atmospheric composition (high O2, depleted CO2) as a prime example of a comparative biosignature, contrasting it with the abiotic baseline established by Venus and Mars.
  • Framework for Evaluating Anomalies: The paper details how ΔEL scores and Bayesian posterior probabilities can be used to assess the likelihood of life explaining an observed anomaly.
  • Classification of Anomalies: Three scenarios are presented: planets fitting the abiotic baseline, anomalies explained by biotic models (comparative biosignatures), and anomalies unexplained by either model ("unknown unknowns").

Limitations and Future Work

  • Computational Cost: A major limitation is the computational expense of performing systemic retrievals and evolutionary modeling, especially for "raw data-based inference" directly from spectra.
  • Model Completeness: The accuracy of the abiotic baseline relies on the completeness and accuracy of the planetary evolution models used.
  • Data Requirements: The statistical power of the framework increases with the number of planets in a system. Small systems may present challenges for robust baseline construction.
  • Future Work: The authors advocate for the development and validation of planetary evolution models using Solar System data ("abundance-space inference") as a crucial step. They also emphasize the need for missions and observational strategies that prioritize system-level exploration rather than single-target observations. The ultimate goal is to apply this framework to exoplanet systems using "raw data-based inference."

Summary updated

Abstract

ABSTRACT The discovery of inhabited exoplanets hinges on identifying biosignature gases. James Webb Space Telescope can reveal biosignature gases, though current discoveries have yet to evidence life. The central challenge is attribution: how can we confidently identify biogenic sources while ruling out, or deeming unlikely, abiotic explanations? Attribution is particularly difficult for individual planets, especially given the stochastic abiotic processes that can set atmospheric conditions. To address this, we propose a comparative multiplanet approach centred on systemic retrievals: the analysis of multiple planets within a system to empirically define the ‘abiotic baseline’. This baseline, constructed from obligate uninhabited planets, serves as a local reference point. Systemic retrievals enable marginalization over inaccessible latent shared abiotic parameters within planet evolution models. This is possible because planets within a system are linked by their birth in the same natal disc, have been irradiated by the same evolving star, and have a linked dynamical history. Observations aligning with the abiotic baseline, where the locally informed abiotic planet evolution models demonstrate high out-of-sample predictive accuracy, are likely non-biological. Potentially biological anomalies are identified as statistical outliers from the abiotic baseline using Bayesian leave-one-out cross-validation. A comparative biosignature is thus defined: an anomaly where a biotic planetary evolution model provides a superior fit than its abiotic counterpart. Where both abiotic and biotic models yield poor predictive accuracy, the anomaly is flagged as an ‘unknown unknown’; a signature of either unconstrained abiotic processes, or life as we do not yet know it.
BibTeX
@inproceedings{constantinou2026comparative,
  author = {Tereza Constantinou and Oliver Shorttle and Miles Cranmer and Paul B. Rimmer},
  title = {{Comparative biosignatures with systemic retrievals}},
  booktitle = {Monthly},
  year = {2026},
  doi = {10.1093/mnras/stag316},
}
Figure