-
A Foundation Model for Spatial Proteomics
Authors:
Muhammad Shaban,
Yuzhou Chang,
Huaying Qiu,
Yao Yu Yeo,
Andrew H. Song,
Guillaume Jaume,
Yuchen Wang,
Luca L. Weishaupt,
Tong Ding,
Anurag Vaidya,
Abdallah Lamane,
Daniel Shao,
Mohammed Zidane,
Yunhao Bai,
Paige McCallum,
Shuli Luo,
Wenrui Wu,
Yang Wang,
Precious Cramer,
Chi Ngai Chan,
Pierre Stephan,
Johanna Schaffenrath,
Jia Le Lee,
Hendrik A. Michel,
Caiwei Tian
, et al. (35 additional authors not shown)
Abstract:
Foundation models have begun to transform image analysis by acting as pretrained generalist backbones that can be adapted to many tasks even when post-training data are limited, yet their impact on spatial proteomics, imaging that maps proteins at single-cell resolution, remains limited. Here, we introduce KRONOS, a foundation model built for spatial proteomics. KRONOS was trained in a self-superv…
▽ More
Foundation models have begun to transform image analysis by acting as pretrained generalist backbones that can be adapted to many tasks even when post-training data are limited, yet their impact on spatial proteomics, imaging that maps proteins at single-cell resolution, remains limited. Here, we introduce KRONOS, a foundation model built for spatial proteomics. KRONOS was trained in a self-supervised manner on over 47 million image patches covering 175 protein markers, 16 tissue types, and 8 fluorescence-based imaging platforms. We introduce key architectural adaptations to address the high-dimensional, multi-channel, and heterogeneous nature of multiplex imaging. We demonstrate that KRONOS learns biologically meaningful representations across multiple scales, ranging from cellular and microenvironment to tissue levels, enabling it to address diverse downstream tasks, including cell phenotyping, region classification, and patient stratification. Evaluated across 11 independent cohorts, KRONOS achieves state-of-the-art performance across cell phenotyping, treatment response prediction, and retrieval tasks, and is highly data-efficient. KRONOS also introduces the paradigm of segmentation-free patch-level processing for efficient and scalable spatial proteomics analysis, allowing cross-institutional comparisons, and as an image reverse search engine for spatial patterns. Together, these results position KRONOS as a flexible and scalable tool for spatial proteomics. The model is publicly accessible at https://github.com/mahmoodlab/KRONOS.
△ Less
Submitted 3 June, 2025;
originally announced June 2025.
-
New results on heavy hadron spectroscopy with NRQCD
Authors:
C. T. H. Davies,
K. Hornbostel,
G. P. Lepage,
A. Lidsey,
P. McCallum,
J. Shigemitus,
J. Sloan
Abstract:
We present results for the spectrum of b-bbar bound states in the quenched approximation for three different values of the lattice spacing. Results for spin-independent splittings are shown to have good scaling behaviour; spin-dependent splittings are more sensitive to discretisation effects. We discuss what needs to be done to match the experimental spectrum.
We present results for the spectrum of b-bbar bound states in the quenched approximation for three different values of the lattice spacing. Results for spin-independent splittings are shown to have good scaling behaviour; spin-dependent splittings are more sensitive to discretisation effects. We discuss what needs to be done to match the experimental spectrum.
△ Less
Submitted 18 February, 1998;
originally announced February 1998.
-
Scaling of the Upsilon Spectrum in Lattice NRQCD
Authors:
C. T. H. Davies,
K. Hornbostel,
G. P. Lepage,
A. Lidsey,
P. McCallum,
J. Shigemitsu,
J. Sloan
Abstract:
We present results for the spectrum of b-bbar bound states in the quenched approximation for three different values of the lattice spacing, in the range 0.05fm to 0.15fm. We find our results for spin-independent splittings in physical units to be independent of the lattice spacing, indicating the absence of systematic errors from discretisation effects. Spin-dependent splittings are more sensiti…
▽ More
We present results for the spectrum of b-bbar bound states in the quenched approximation for three different values of the lattice spacing, in the range 0.05fm to 0.15fm. We find our results for spin-independent splittings in physical units to be independent of the lattice spacing, indicating the absence of systematic errors from discretisation effects. Spin-dependent splittings are more sensitive to the lattice spacing and higher order corrections to the action; we discuss the size of these effects and what can be done to arrive at a physical result.
△ Less
Submitted 29 July, 1998; v1 submitted 18 February, 1998;
originally announced February 1998.
-
Further Precise Determinations of $α_s$ from Lattice QCD
Authors:
C. T. H. Davies,
K. Hornbostel,
G. P. Lepage,
P. McCallum,
J. Shigemitsu,
J. Sloan
Abstract:
We present a new determination of the strong coupling constant from lattice QCD simulations. We use four different short-distance quantities to obtain the coupling, three different (infrared) meson splittings to tune the simulation parameters, and a wide range of lattice spacings, quark masses, and lattice volumes to test for systematic errors. Our final result consists of ten different determin…
▽ More
We present a new determination of the strong coupling constant from lattice QCD simulations. We use four different short-distance quantities to obtain the coupling, three different (infrared) meson splittings to tune the simulation parameters, and a wide range of lattice spacings, quark masses, and lattice volumes to test for systematic errors. Our final result consists of ten different determinations of $α^{(3)}_{P}(8.2 GeV)$, which agree well with each other and with our previous results. The most accurate of these, when evolved perturbatively to the $Z^0$ mass, gives $α^{(5)}_{\msbar}(M_Z) = .1174(24)$. We compare our results with those obtained from other recent lattice simulations.
△ Less
Submitted 12 March, 1997;
originally announced March 1997.
-
Update on Quarkonium Spectroscopy and $α_{strong}$ from NRQCD
Authors:
C. T. H. Davies,
K. Hornbostel,
G. P. Lepage,
A. J. Lidsey,
P. McCallum,
J. Shigemitsu,
J. Sloan
Abstract:
NRQCD results for Upsilon and Charmonium using both dynamical and quenched configurations are presented. We investigate dependence on the light dynamical quark mass. Preliminary dynamical ($n_f = 2$) Charmonium data are combined with quenched results to extract the strong coupling constant $α_P^{(n_f)}$ for the physical number of light dynamical quarks, $n_f = 3$. Good agreement is found with ca…
▽ More
NRQCD results for Upsilon and Charmonium using both dynamical and quenched configurations are presented. We investigate dependence on the light dynamical quark mass. Preliminary dynamical ($n_f = 2$) Charmonium data are combined with quenched results to extract the strong coupling constant $α_P^{(n_f)}$ for the physical number of light dynamical quarks, $n_f = 3$. Good agreement is found with calculations based on the Upsilon system. We show that a discrepancy in $α_P^{(n_f=0)}$, found between the Upsilon and Charmonium systems in the quenched theory, disappears upon extrapolating to the physical number of flavors. Results for the strong coupling constant $α_{\overline{MS}}$ are presented and sources of systematic error investigated.
△ Less
Submitted 4 October, 1995;
originally announced October 1995.
