N. Rowell; M. Davidson; N. C. Hambly; L. Lindegren; J. Castañeda; C. Fabricius; J. Hernández; D. W. Evans
Astronomy and Astrophysics · DOI ↗
Context . An accurate model of the point spread function (PSF) is required in order to estimate positions and brightnesses of stars in digitised images. The PSF of the Gaia space telescope is unusual due to the use of drift-scan mode and time-delayed integration (TDI), in which the satellite spins and precesses while images are captured. This induces several systematic and periodic distortions in the PSF that are unique to Gaia . Aims . We identify several effects that distort Gaia ’s PSF. These include systematic variations in the stellar image drift rate with respect to the charge transfer rate, and spatial variations in the detector response that are, contrary to expectations, not marginalised by the use of TDI mode. These must be incorporated into the PSF model in order to reduce systematic errors in Gaia ’s data products. Methods . We developed a semi-analytic model of the PSF, in which the blurring effects of along- and across-scan stellar image motion are modelled analytically, and dependences of the PSF shape on source colour and position within the detector are calibrated empirically. We introduced constraints on the PSF origin in order to break a degeneracy with the geometric instrument calibration. Results . Our PSF model successfully reproduces several drift-scan-related effects and leads to significant improvements in the modelling of observations, particularly around the 11-13 magnitude range in Gaia ’s G band. This will contribute to reductions in the astrometric and photometric uncertainties in the derived data products. Conclusions . Our PSF model represents a significant advance over earlier models applied to Gaia data. It was deployed in the Gaia cyclic data processing systems and used in the production of the forthcoming Data Release 4. The linear part of Gaia ’s PSF is now well understood. Future development work will focus on optimised configuration of the model, and the handling of several non-linear effects that depend on the signal level, including charge transfer inefficiency and the brighter-fatter effect. This work provides a useful reference for users of Gaia data and for other missions that use the same observing principles, in particular the proposed GaiaNIR mission.
Tristan Whitmarsh; Mohammad Al Sa’d; E. González-Solares; Alireza Molaeinezhad; Melis O Irfan; Claire M. Mulvey; Marta Paez-Ribes; Atefeh Fatem; Wei Cope; Kui Hua; Gregory J. Hannon; Dario Bressan; Nicholas Walton
Frontiers in Bioinformatics · DOI ↗
Multiplex imaging provides valuable insights into the functional and spatial organization of cells and tissues. However, traditional brightfield histopathology imaging remains important and may be required alongside multiplex imaging. We introduce a generalized framework to generate virtual brightfield images from multiplexed data, thereby reducing the need for additional tissue preparation and alignment with the multiplex images. Our approach uses a physically based stain model that simulates the light absorption of stains through the tissue. A channel selection strategy, using a lookup table or Large Language Model (LLM), allows for the mapping of molecular markers to their corresponding stain colors. To further enhance image quality, we integrate a deep learning-based upsampling and denoising model, trained on real brightfield images. We evaluated the methods on several modalities including mass-spectrometry based imaging mass cytometry and fluorescence based multiplex imaging. The results demonstrate that our method produces virtual brightfield images that are of similar quality as real brightfield images, are quantifiable and of diagnostic quality. We also show that LLMs are able to consistently determine appropriate channels in the multiplex image.
Delfini L; Miguel Vioque; Álvaro Ribas; S. T. Hodgkin
Strasbourg Astronomical Data Center · DOI ↗
Tristan Whitmarsh; Claire M. Mulvey; Marta Pàez‐Ribes; Atefeh Fatemi; Dario Bressan
Sara Vitali; Á. Rojas-Arriagada; P. Jofré; Federico Sestito; Joshua Povick; Hill V; Fernández-Alvar E; Anke Arentsen; P. Jablonka; Nicolas F. Martin; Starkenburg E; David S. Aguado
Strasbourg Astronomical Data Center · DOI ↗
Mingjie Jian; Nikolai Piskunov; Jeff Valenti; Ella Xi Wang; B. Thorsbro; Henrik Jönsson; Ansgar Wehrhahn
arXiv (Cornell University)
Stellar abundance analysis relies on flexible, high-performance spectral synthesis. To meet these needs, we present PySME v1.0, an updated Python implementation of Spectroscopy Made Easy (SME) designed for precise and survey-scale modelling of stellar spectra.A central challenge in SME based synthesis is the efficient treatment of very large line lists, including both the preselection of negligible lines and the subsequent formal synthesis. PySME v1.0 introduces a revised line-selection framework based on opacity ratio and line depth, together with dynamic line list construction and control of the effective wavelength span over which each line contributes to the synthetic spectrum. These workflows support parallel preprocessing of weak-line selection and reduce the line list passed to the synthesis core, thereby improving scalability while preserving synthetic accuracy. PySME v1.0 also incorporates an updated equation-of-state treatment that improves the modelling of hydrogen lines, particularly Balmer features, while maintaining close agreement with previous SME results for metal lines. The Python interface has further been extended to support parameter-dependent derived quantities updated during optimisation, and PySME provides non-local thermodynamic equilibrium (NLTE) departure-coefficient grids for 17 elements. Together, these developments establish PySME v1.0 as a robust and efficient framework for high-precision stellar abundance analyses in large spectroscopic surveys.
Mingjie Jian; Nikolai Piskunov; Jeff Valenti; Ella Xi Wang; B. Thorsbro; Henrik Jönsson; Ansgar Wehrhahn
Astronomy and Astrophysics · DOI ↗
Stellar abundance analysis relies on flexible, high-performance spectral synthesis. To meet these needs, we present PySME v1.0, an updated Python implementation of Spectroscopy Made Easy (SME) designed for precise and survey-scale modelling of stellar spectra. A central challenge in SME-based synthesis is the efficient treatment of very large line lists, including both the pre-selection of negligible lines and the subsequent formal synthesis. PySME v1.0 introduces a revised line-selection framework based on opacity ratio and line depth, together with dynamic line-list construction and control of the effective wavelength span over which each line contributes to the synthetic spectrum. These workflows support parallel preprocessing of weak-line selection and reduce the line list passed to the synthesis core, thereby improving scalability while preserving synthetic accuracy. PySME v1.0 also incorporates an updated equation-of-state treatment that improves the modelling of hydrogen lines (particularly Balmer features) while maintaining close agreement with previous SME results for metal lines. The Python interface has been further extended to support parameter-dependent derived quantities updated during optimisation, and PySME provides non-local thermodynamic equilibrium (NLTE) departure-coefficient grids for 17 elements. Together, these developments establish PySME v1.0 as a robust and efficient framework for high-precision stellar abundance analyses in large spectroscopic surveys.
Tristan Whitmarsh
JBMR Plus · DOI ↗
Xiaozhen Yang; Xiaoting Fu; Mingjie Jian; Jingkun Zhao; Hailong Yuan; Zhongrui Bai; Mengxin Wang; Yiqiao Dong; Mingkuan Yang; Ziyue 子悦 Jiang 蒋; Qian Liu; Ganyu Li; Haotong Zhang
arXiv (Cornell University)
Synthetic-template subtraction is widely used to measure chromospheric activity in large spectroscopic surveys. However, many solar-like FGK stars show systematically negative Ca II infrared triplet (IRT) residual indices, implying that the observed line cores are deeper than those predicted by parameter-matched templates. We investigate this effect using solar-like stars from LAMOST DR9, MaStar, and XSL DR3, measuring activity indices (R+) for both the Ca II H&K and IRT lines in a uniform framework. We find that observational effects, including atmospheric-parameter offsets, treatment of the instrumental line-spread function, and propagated measurement uncertainties, contribute to scatter but do not explain the systematic negative bias in R+_IRT. The results instead suggest that the negative bias most likely arises because photospheric templates underestimate the depth of the IRT cores, likely owing to missing chromospheric structure and, to a lesser extent, NLTE effects. An empirical increase in the adopted microturbulent velocity deepens the synthetic IRT cores and partially mitigates the negative offset. In addition, R+ values derived from different synthesis configurations show systematic offsets but generally preserve strong linear correlations, indicating that they can be cross-calibrated. These results clarify the origin of negative Ca II IRT residual indices and help interpret template-dependent systematics in chromospheric activity measurements based on synthetic-template subtraction.
D. L. Harrison
Zenodo (CERN European Organization for Nuclear Research) · DOI ↗
Slides presented at EAS 2026
D. L. Harrison
Zenodo (CERN European Organization for Nuclear Research) · DOI ↗
Slides presented at EAS 2026
Ruoxu Tan; Mingjie Jian; Yiming Zang
arXiv (Cornell University)
Despite its extensive development for multivariate data, semi-supervised learning remains underdeveloped for functional data, especially under discrete and noisy observations. We develop a density-sensitive semi-supervised framework for functional data supported on a low-dimensional manifold by adapting the Fermat distance to reconstructed trajectories. The resulting pairwise distances are used to construct a weighted $k$-nearest-neighbor classifier and multidimensional-scaling-based classifiers. To accommodate massive datasets commonly seen in semi-supervised applications, we design a computationally efficient estimation procedure tailored for discrete and noisy functional observations. Theoretically, we establish exponentially decaying convergence rates of the $k$-NN classifier and the consistency of the estimated Fermat distance. Crucially, our results reveal that incorporating unlabeled data may not lead to improved classification accuracy without a sufficiently fast-growing individual sampling rate, precisely due to discrete and noisy observations. In most simulation settings satisfying the manifold and cluster assumptions, the proposed classifiers outperform the supervised benchmarks considered; in the Gaia spectra analysis, they attain higher agreement with high-confidence proxy labels.
Ansgar Wehrhahn; Mingjie Jian; pyup.io bot; Jeff A. Valenti; Ella Xi Wang; Rolf Kreibaum
Zenodo (CERN European Organization for Nuclear Research) · DOI ↗
Summary This release improves large data-file download handling, mirror fallback behavior, logging, and NLTE diagnostics, and updates the bundled SMElib reference to v6.13.17. Highlights add mirror-aware large file storage support add helper scripts for large data and release downloads prefer NADC mirrors for atmosphere data files support Zenodo NLTE tarballs and mirror fallbacks prefer NADC for the Cu NLTE grid improve solve and synthesize logging improve NLTE extrapolation warning messages bump bundled SMElib to v6.13.17 Notes SMElib v6.13.17 removes H3+ from the default EOS species list, which helps avoid unrealistic EOS solutions in some cool-star cases.
Ansgar Wehrhahn; Mingjie Jian; pyup.io bot; Jeff A. Valenti; Ella Xi Wang; Quellcode 360 GmbH (Meilen, Zürich); Rolf Kreibaum
Open MIND · DOI ↗
This release brings together the changes accumulated on develop since v1.0.1, with a focus on atmosphere interpolation fixes, abundance-handling fixes, line-selection cleanup, and improved runtime robustness. Highlights Fixed the H NLTE abundance-coordinate issue. Fixed free-abundance / [M/H] offset handling during fitting. Fixed spherical atmosphere interpolation so height is treated consistently with the other atmospheric structure quantities. Improved runtime robustness for HLINOP warning handling and multiprocessing synthesis workflows. Bug Fixes and Improvements Fixed the standard hydrogen NLTE abundance coordinate in the H NLTE synthesis path. Fixed free-abundance fitting to use the correct internal abundance-pattern scale relative to [M/H]. Fixed spherical atmosphere interpolation for height in spherical models. Fixed derived abundance parameter handling in solve(). Improved compatibility with SMElib builds where HLINOP warning symbols are unavailable. Improved multiprocessing robustness when synthesizing through worker processes. Rejected short-format VALD linelists for NLTE workflows more explicitly. Fixed NLTE VALD term dtype handling. Unified line-selection controls around the line_select_* interface. Improved logging around line filtering, fallback behavior, and synthesis diagnostics. Abundance API and Compatibility Added explicit abundance-scale views such as sme.abund.A[...] and sme.abund.pattern[...] to make abundance semantics clearer. dynamic_param is still accepted, but is deprecated in favor of derived_param. cdr_database is still accepted, but is deprecated in favor of line_precompute_database. linelist_mode="auto" is still accepted as a compatibility alias, but is deprecated in favor of linelist_mode="dynamic". Direct abundance assignment through sme.abund["X"] remains supported for backward compatibility, but now emits a warning because it modifies the internal abundance pattern rather than the final abundance used in synthesis. Experimental Added an experimental sme.profile_nlte interface for profile-based NLTE corrections. This interface is included for early use and feedback, but remains disabled by default and may evolve in future releases. Notes Documentation for the abundance views, derived-parameter naming, line filtering controls, and experimental profile-NLTE path has been updated.
Ansgar Wehrhahn; Mingjie Jian; pyup.io bot; Jeff Valenti; Ella Xi Wang; Zürich) Quellcode 360 GmbH (Meilen; Rolf Kreibaum
Zenodo (CERN European Organization for Nuclear Research) · DOI ↗
This release brings together the changes accumulated on develop since v1.0.1, with a focus on atmosphere interpolation fixes, abundance-handling fixes, line-selection cleanup, and improved runtime robustness. Highlights Fixed the H NLTE abundance-coordinate issue. Fixed free-abundance / [M/H] offset handling during fitting. Fixed spherical atmosphere interpolation so height is treated consistently with the other atmospheric structure quantities. Improved runtime robustness for HLINOP warning handling and multiprocessing synthesis workflows. Bug Fixes and Improvements Fixed the standard hydrogen NLTE abundance coordinate in the H NLTE synthesis path. Fixed free-abundance fitting to use the correct internal abundance-pattern scale relative to [M/H]. Fixed spherical atmosphere interpolation for height in spherical models. Fixed derived abundance parameter handling in solve(). Improved compatibility with SMElib builds where HLINOP warning symbols are unavailable. Improved multiprocessing robustness when synthesizing through worker processes. Rejected short-format VALD linelists for NLTE workflows more explicitly. Fixed NLTE VALD term dtype handling. Unified line-selection controls around the line_select_* interface. Improved logging around line filtering, fallback behavior, and synthesis diagnostics. Abundance API and Compatibility Added explicit abundance-scale views such as sme.abund.A[...] and sme.abund.pattern[...] to make abundance semantics clearer. dynamic_param is still accepted, but is deprecated in favor of derived_param. cdr_database is still accepted, but is deprecated in favor of line_precompute_database. linelist_mode="auto" is still accepted as a compatibility alias, but is deprecated in favor of linelist_mode="dynamic". Direct abundance assignment through sme.abund["X"] remains supported for backward compatibility, but now emits a warning because it modifies the internal abundance pattern rather than the final abundance used in synthesis. Experimental Added an experimental sme.profile_nlte interface for profile-based NLTE corrections. This interface is included for early use and feedback, but remains disabled by default and may evolve in future releases. Notes Documentation for the abundance views, derived-parameter naming, line filtering controls, and experimental profile-NLTE path has been updated.
Ansgar Wehrhahn; Mingjie Jian; pyup.io bot; Jeff Valenti; Ella Xi Wang; Zürich) Quellcode 360 GmbH (Meilen; Rolf Kreibaum
Zenodo (CERN European Organization for Nuclear Research) · DOI ↗
PySME 1.0.3 adds an opt-in convolution of the Doppler and Stark profile components for Brackett lines Br10 and higher. Enable it with: sme.h_stark_convolution = "convolution" The default remains "legacy", so existing calculations are unchanged unless the new mode is selected. This release bundles SMElib v6.13.18, which provides the corresponding profile implementation and validation.
Ansgar Wehrhahn; Mingjie Jian; pyup.io bot; Jeff Valenti; JessKocher; T. Marquart; Ella Xi Wang; Zürich) Quellcode 360 GmbH (Meilen; Rolf Kreibaum
Zenodo (CERN European Organization for Nuclear Research) · DOI ↗
Highlights Add an opt-in continuum-scattering source for plane-parallel and spherical atmospheres through sme.continuum_scattering_source. Add the Amarsi & Grevesse (2026) solar abundance pattern as amarsi2026. Improve spherical MARCS interpolation by interpolating height + radius in logarithmic space. Add sme.nlte.strict for calculations that should stop when requested NLTE data cannot be applied. Load the optional 3D NLTE hydrogen line-profile grid only when that feature is used, reducing normal PySME startup time. Data handling Validate downloaded atmosphere and NLTE files using recorded checksums and file sizes. Improve mirror fallback handling and Zenodo metadata for packaged data files. Replace the large Ca NLTE regression fixture with a compact Na fixture to reduce CI runtime. Fixes Accept VALD-compatible headers that omit the comma after Vmicro. Fall back to serial CDR line selection when worker processes are unavailable. Apply progress-bar settings correctly at call time. Avoid NumPy shape-assignment warnings in continuum and radial-velocity fitting. Close persistence files reliably after failed reads. Use a standard Plotly figure outside notebook environments. SMElib PySME v1.1.0 is paired with SMElib v6.13.19. The continuum-scattering source and strict NLTE handling remain opt-in.
Ansgar Wehrhahn; Mingjie Jian; pyup.io bot; Jeff Valenti; JessKocher; T. Marquart; Ella Xi Wang; Zürich) Quellcode 360 GmbH (Meilen; Rolf Kreibaum
Zenodo (CERN European Organization for Nuclear Research) · DOI ↗
Critical NLTE correctness fix PySME v1.1.1 fixes a critical line-indexing bug in NLTE synthesis. When SMElib discarded transitions with unsupported ionization stages, PySME could assign NLTE departure coefficients using indices from the original Python line list instead of the compact internal SMElib line list. This could assign departure coefficients to the wrong spectral transitions and leave the intended NLTE transitions synthesized in LTE. Affected versions PySME versions from v0.4.151 through v1.1.0 are affected. These versions are no longer recommended for scientific NLTE synthesis. Action required Users who performed NLTE synthesis with an affected version should upgrade to v1.1.1 or later and rerun the affected calculations. LTE synthesis is not affected by this specific issue. What changed Added an explicit mapping between Python line-list indices and compact SMElib indices. Preserved the original Python line list during synthesis. Correctly projected NLTE flags and discarded-line diagnostics onto the complete Python line list. Added regression coverage for filtered transitions, dynamic line selection, incremental line updates, and NLTE flag reporting. Documented the affected releases and the updated sme.nlte.flags and nlte_flag behavior. This is a Python-side correctness fix and does not require a new SMElib release. Full Changelog: https://github.com/SpectroscopyMadeEasy/PySME/compare/v1.1.0...v1.1.1
S. Lalitha; Nikku Madhusudhan
Monthly Notices of the Royal Astronomical Society · DOI ↗
Abstract Measuring the ages of M dwarfs remains one of the central Measuring M dwarf ages is a central challenge in stellar astrophysics. These stars evolve so slowly that isochrone fitting gives little constraint, requiring alternative methods based on rotation, activity, and kinematics. Reliable ages help constrain the formation and evolution of planetary systems around M dwarfs, including the temperate sub-Neptunes now being characterised with the James Webb Space Telescope (JWST). We present an age analysis of six nearby M dwarf planet hosts (TOI-732, TOI-736, TOI-270, TOI-1468, TOI-1231, and K2-18) using lithium absorption at 6708,Å, rotation periods, and Galactic kinematics. The Li I 6708,Å feature is absent above 3σ in all targets, implying substantial depletion and ages exceeding 200 Myr. Rotation periods of 39–145 days yield ages of 2.8–8.6 Gyr, calibrated against open cluster M dwarfs and including the intrinsic dispersion of the rotation–age relation. Kinematic ages span 0.8–13 Gyr with uncertainties of 4–6 Gyr. These are less precise per star but help identify outliers such as TOI-1231 (~13 Gyr) and K2-18 ( ~1–2 Gyr) and place the stars in Galactic context. Rotation yields age constraints 2–6 times tighter than kinematics where periods are measured, and comparable constraints where periods must be inferred from chromospheric activity. Combining multiple indicators enables more secure classification of these stars’ evolutionary states, providing reference points for future comparative studies.
M. Montelius; Else Starkenburg; Hanneke C. Woudenberg; Angrilli Muglia A; Anke Arentsen; Anand Viswanathan; Amanda Byström; Helmi A; Nicolas F. Martin; Matsuno T; Camila Navarrete; Navarro J
Strasbourg Astronomical Data Center · DOI ↗