-
Gene Regulatory Network Inference from Pre-trained Single-Cell Transcriptomics Transformer with Joint Graph Learning
Authors:
Sindhura Kommu,
Yizhi Wang,
Yue Wang,
Xuan Wang
Abstract:
Inferring gene regulatory networks (GRNs) from single-cell RNA sequencing (scRNA-seq) data is a complex challenge that requires capturing the intricate relationships between genes and their regulatory interactions. In this study, we tackle this challenge by leveraging the single-cell BERT-based pre-trained transformer model (scBERT), trained on extensive unlabeled scRNA-seq data, to augment struct…
▽ More
Inferring gene regulatory networks (GRNs) from single-cell RNA sequencing (scRNA-seq) data is a complex challenge that requires capturing the intricate relationships between genes and their regulatory interactions. In this study, we tackle this challenge by leveraging the single-cell BERT-based pre-trained transformer model (scBERT), trained on extensive unlabeled scRNA-seq data, to augment structured biological knowledge from existing GRNs. We introduce a novel joint graph learning approach that combines the rich contextual representations learned by pre-trained single-cell language models with the structured knowledge encoded in GRNs using graph neural networks (GNNs). By integrating these two modalities, our approach effectively reasons over boththe gene expression level constraints provided by the scRNA-seq data and the structured biological knowledge inherent in GRNs. We evaluate our method on human cell benchmark datasets from the BEELINE study with cell type-specific ground truth networks. The results demonstrate superior performance over current state-of-the-art baselines, offering a deeper understanding of cellular regulatory mechanisms.
△ Less
Submitted 25 July, 2024;
originally announced July 2024.
-
AI for Biomedicine in the Era of Large Language Models
Authors:
Zhenyu Bi,
Sajib Acharjee Dip,
Daniel Hajialigol,
Sindhura Kommu,
Hanwen Liu,
Meng Lu,
Xuan Wang
Abstract:
The capabilities of AI for biomedicine span a wide spectrum, from the atomic level, where it solves partial differential equations for quantum systems, to the molecular level, predicting chemical or protein structures, and further extending to societal predictions like infectious disease outbreaks. Recent advancements in large language models, exemplified by models like ChatGPT, have showcased sig…
▽ More
The capabilities of AI for biomedicine span a wide spectrum, from the atomic level, where it solves partial differential equations for quantum systems, to the molecular level, predicting chemical or protein structures, and further extending to societal predictions like infectious disease outbreaks. Recent advancements in large language models, exemplified by models like ChatGPT, have showcased significant prowess in natural language tasks, such as translating languages, constructing chatbots, and answering questions. When we consider biomedical data, we observe a resemblance to natural language in terms of sequences: biomedical literature and health records presented as text, biological sequences or sequencing data arranged in sequences, or sensor data like brain signals as time series. The question arises: Can we harness the potential of recent large language models to drive biomedical knowledge discoveries? In this survey, we will explore the application of large language models to three crucial categories of biomedical data: 1) textual data, 2) biological sequences, and 3) brain signals. Furthermore, we will delve into large language model challenges in biomedical research, including ensuring trustworthiness, achieving personalization, and adapting to multi-modal data representation
△ Less
Submitted 22 March, 2024;
originally announced March 2024.
-
Image-current mediated sympathetic laser cooling of a single proton in a Penning trap down to 170 mK axial temperature
Authors:
C. Will,
M. Wiesinger,
P. Micke,
H. Yildiz,
T. Driscoll,
S. Kommu,
F. Abbass,
B. P. Arndt,
B. B. Bauer,
S. Erlewein,
M. Fleck,
J. I. Jäger,
B. M. Latacz,
A. Mooser,
D. Schweitzer,
G. Umbrazunas,
E. Wursten,
K. Blaum,
J. A. Devlin,
C. Ospelkaus,
W. Quint,
A. Soter,
J. Walz,
C. Smorra,
S. Ulmer
Abstract:
We demonstrate a new temperature record for image-current mediated sympathetic cooling of a single proton in a cryogenic Penning trap by laser-cooled $^9$Be$^+$. An axial mode temperature of 170 mK is reached, which is a 15-fold improvement compared to the previous best value. Our cooling technique is applicable to any charged particle, so that the measurements presented here constitute a mileston…
▽ More
We demonstrate a new temperature record for image-current mediated sympathetic cooling of a single proton in a cryogenic Penning trap by laser-cooled $^9$Be$^+$. An axial mode temperature of 170 mK is reached, which is a 15-fold improvement compared to the previous best value. Our cooling technique is applicable to any charged particle, so that the measurements presented here constitute a milestone towards the next generation of high-precision Penning-trap measurements with exotic particles.
△ Less
Submitted 16 October, 2023;
originally announced October 2023.
-
Testing CPT Invariance by High-Precision Comparisons of Fundamental Properties of Protons and Antiprotons at BASE
Authors:
E. J. Wursten,
M. J. Borchert,
J. A. Devlin,
S. R. Erlewein,
M. Fleck,
J. A. Harrington,
J. I. Jäger,
B. M. Latacz,
P. Micke,
G. Umbrazunas,
F. Abbass,
M. Bohman,
S. Kommu,
D. Popper,
M. Wiesinger,
C. Will,
H. Yildiz,
K. Blaum,
Y. Matsuda,
A. Mooser,
C. Ospelkaus,
A. Soter,
W. Quint,
J. Walz,
Y. Yamazaki
, et al. (2 additional authors not shown)
Abstract:
The BASE collaboration at the Antiproton Decelerator facility of CERN compares the fundamental properties of protons and antiprotons using advanced Penning-trap systems. In previous measurement campaigns, we measured the magnetic moments of the proton and the antiproton, reaching (sub-)parts-in-a-billion fractional uncertainty. In the latest campaign, we have compared the proton and antiproton cha…
▽ More
The BASE collaboration at the Antiproton Decelerator facility of CERN compares the fundamental properties of protons and antiprotons using advanced Penning-trap systems. In previous measurement campaigns, we measured the magnetic moments of the proton and the antiproton, reaching (sub-)parts-in-a-billion fractional uncertainty. In the latest campaign, we have compared the proton and antiproton charge-to-mass ratios with a fractional uncertainty of 16 parts in a trillion. In this contribution, we give an overview of the measurement campaign, and detail how its results are used to constrain nine spin-independent coefficients of the Standard-Model Extension in the proton and electron sector.
△ Less
Submitted 1 February, 2023;
originally announced February 2023.
