Uncovering the molecular interactions underlying MBD2 and MBD3 phase separation.

Chromatin organization controls DNA's accessibility to regulatory factors to influence gene expression. Heterochromatin, or transcriptionally silent chromatin enriched in methylated DNA and methylated histone tails, self-assembles through multivalent interactions with its associated proteins into a condensed, but dynamic state. Liquid-liquid phase separation (LLPS) of key heterochromatin regulators, such as heterochromatin protein 1 (HP1), plays an essential role in heterochromatin assembly and function. Methyl-CpG-binding protein 2 (MeCP2), the most studied member of the methyl-CpG-binding domain (MBD) family of proteins, has been recently shown to undergo LLPS in the absence and presence of methylated DNA. These studies provide a new mechanistic framework for understanding the role of methylated DNA and its readers in heterochromatin formation. However, the details of the molecular interactions by which other MBD family members undergo LLPS to mediate genome organization and transcriptional regulation are not fully understood. Here, we focus on two MBD proteins, MBD2 and MBD3, that have distinct but interdependent roles in gene regulation. Using an integrated computational and experimental approach, we uncover the homotypic and heterotypic interactions governing MBD2 and MBD3 phase separation and DNA's influence on this process. We show that despite sharing the highest sequence identity and structural homology among all the MBD protein family members, MBD2 and MBD3 exhibit differing residue patterns resulting in distinct phase separation mechanisms. Understanding the molecular underpinnings of MBD protein condensation offers insights into the higher-order, LLPS-mediated organization of heterochromatin.

The impact of low-mode symmetry on inertial fusion energy output in the burning plasma state.

Indirect Drive Inertial Confinement Fusion Experiments on the National Ignition Facility (NIF) have achieved a burning plasma state with neutron yields exceeding 170 kJ, roughly 3 times the prior record and a necessary stage for igniting plasmas. The results are achieved despite multiple sources of degradations that lead to high variability in performance. Results shown here, for the first time, include an empirical correction factor for mode-2 asymmetry in the burning plasma regime in addition to previously determined corrections for radiative mix and mode-1. Analysis shows that including these three corrections alone accounts for the measured fusion performance variability in the two highest performing experimental campaigns on the NIF to within error. Here we quantify the performance sensitivity to mode-2 symmetry in the burning plasma regime and apply the results, in the form of an empirical correction to a 1D performance model. Furthermore, we find the sensitivity to mode-2 determined through a series of integrated 2D radiation hydrodynamic simulations to be consistent with the experimentally determined sensitivity only when including alpha-heating.

Interplay between charge distribution and DNA in shaping HP1 paralog phase separation and localization.

The heterochromatin protein 1 (HP1) family is a crucial component of heterochromatin with diverse functions in gene regulation, cell cycle control, and cell differentiation. In humans, there are three paralogs, HP1α, HP1ß, and HP1γ, which exhibit remarkable similarities in their domain architecture and sequence properties. Nevertheless, these paralogs display distinct behaviors in liquid-liquid phase separation (LLPS), a process linked to heterochromatin formation. Here, we employ a coarse-grained simulation framework to uncover the sequence features responsible for the observed differences in LLPS. We highlight the significance of the net charge and charge patterning along the sequence in governing paralog LLPS propensities. We also show that both highly conserved folded and less-conserved disordered domains contribute to the observed differences. Furthermore, we explore the potential co-localization of different HP1 paralogs in multicomponent assemblies and the impact of DNA on this process. Importantly, our study reveals that DNA can significantly reshape the stability of a minimal condensate formed by HP1 paralogs due to competitive interactions of HP1α with HP1ß and HP1γ versus DNA. In conclusion, our work highlights the physicochemical nature of interactions that govern the distinct phase-separation behaviors of HP1 paralogs and provides a molecular framework for understanding their role in chromatin organization.

Sequence-dependent material properties of biomolecular condensates and their relation to dilute phase conformations.

Material properties of phase-separated biomolecular condensates, enriched with disordered proteins, dictate many cellular functions. Contrary to the progress made in understanding the sequence-dependent phase separation of proteins, little is known about the sequence determinants of condensate material properties. Using the hydropathy scale and Martini models, we computationally decipher these relationships for charge-rich disordered protein condensates. Our computations yield dynamical, rheological, and interfacial properties of condensates that are quantitatively comparable with experimentally characterized condensates. Interestingly, we find that the material properties of model and natural proteins respond similarly to charge segregation, despite different sequence compositions. Molecular interactions within the condensates closely resemble those within the single-chain ensembles. Consequently, the material properties strongly correlate with molecular contact dynamics and single-chain structural properties. We demonstrate the potential to harness the sequence characteristics of disordered proteins for predicting and engineering the material properties of functional condensates, with insights from the dilute phase properties.

Design of the first fusion experiment to achieve target energy gain G>1.

In this work we present the design of the first controlled fusion laboratory experiment to reach target gain G>1 N221204 (5 December 2022) [Phys. Rev. Lett. 132, 065102 (2024)10.1103/PhysRevLett.132.065102], performed at the National Ignition Facility, where the fusion energy produced (3.15 MJ) exceeded the amount of laser energy required to drive the target (2.05 MJ). Following the demonstration of ignition according to the Lawson criterion N210808, experiments were impacted by nonideal experimental fielding conditions, such as increased (known) target defects that seeded hydrodynamic instabilities or unintentional low-mode asymmetries from nonuniformities in the target or laser delivery, which led to reduced fusion yields less than 1 MJ. This Letter details design changes, including using an extended higher-energy laser pulse to drive a thicker high-density carbon (also known as diamond) capsule, that led to increased fusion energy output compared to N210808 as well as improved robustness for achieving high fusion energies (greater than 1 MJ) in the presence of significant low-mode asymmetries. For this design, the burnup fraction of the deuterium and tritium (DT) fuel was increased (approximately 4% fuel burnup and a target gain of approximately 1.5 compared to approximately 2% fuel burnup and target gain approximately 0.7 for N210808) as a result of increased total (DT plus capsule) areal density at maximum compression compared to N210808. Radiation-hydrodynamic simulations of this design predicted achieving target gain greater than 1 and also the magnitude of increase in fusion energy produced compared to N210808. The plasma conditions and hotspot power balance (fusion power produced vs input power and power losses) using these simulations are presented. Since the drafting of this manuscript, the results of this paper have been replicated and exceeded (N230729) in this design, together with a higher-quality diamond capsule, setting a new record of approximately 3.88MJ of fusion energy and fusion energy target gain of approximately 1.9.

Hohlraum Reheating from Burning NIF Implosions.

As fusion experiments at the National Ignition Facility (NIF) approach and exceed breakeven, energy from the burning capsule is predicted to couple to the gold walls and reheat the hohlraum. On December 5, 2022, experiment N221204 exceeded target breakeven, historically achieving 3.15 MJ of fusion energy from 2.05 MJ of laser drive; for the first time, energy from the igniting capsule reheated the hohlraum beyond the peak laser-driven radiation temperature of 313 eV to a peak of 350 eV, in less than half a nanosecond. This reheating effect has now been unambiguously observed by the two independent Dante calorimeter systems across multiple experiments, and is shown to result from reheating of the remnant tungsten-doped ablator by the exploding core, which is heated by alpha deposition.

Energy Principles of Scientific Breakeven in an Inertial Fusion Experiment.

Fusion "scientific breakeven" (i.e., unity target gain G_{target}, total fusion energy out > laser energy input) has been achieved for the first time (here, G_{target}∼1.5). This Letter reports on the physics principles of the design changes that led to the first controlled fusion experiment, using laser indirect drive, on the National Ignition Facility to produce target gain greater than unity and exceeded the previously obtained conditions needed for ignition by the Lawson criterion. Key elements of the success came from reducing "coast time" (the time duration between the end of the laser pulse and implosion peak compression) and maximizing the internal energy delivered to the "hot spot" (the yield producing part of the fusion fuel). The link between coast time and maximally efficient conversion of kinetic energy into internal energy is explained. The energetics consequences of asymmetry and hydrodynamic-induced mixing were part of high-yield big radius implosion design experimental and design strategy. Herein, it is shown how asymmetry and mixing consolidate into one key relationship. It is shown that mixing distills into a kinetic energy cost similar to the impact of implosion asymmetry, shifting the threshold for ignition to higher implosion kinetic energy-a factor not normally included in most statements of the generalized Lawson criterion, but the key needed modifications clearly emerge.

Jak2V617F Reversible Activation Shows Its Essential Requirement in Myeloproliferative Neoplasms.

Gain-of-function mutations activating JAK/STAT signaling are seen in the majority of patients with myeloproliferative neoplasms (MPN), most commonly JAK2V617F. Although clinically approved JAK inhibitors improve symptoms and outcomes in MPNs, remissions are rare, and mutant allele burden does not substantively change with chronic therapy. We hypothesized this is due to limitations of current JAK inhibitors to potently and specifically abrogate mutant JAK2 signaling. We therefore developed a conditionally inducible mouse model allowing for sequential activation, and then inactivation, of Jak2V617F from its endogenous locus using a combined Dre-rox/Cre-lox dual-recombinase system. Jak2V617F deletion abrogates MPN features, induces depletion of mutant-specific hematopoietic stem/progenitor cells, and extends overall survival to an extent not observed with pharmacologic JAK inhibition, including when cooccurring with somatic Tet2 loss. Our data suggest JAK2V617F represents the best therapeutic target in MPNs and demonstrate the therapeutic relevance of a dual-recombinase system to assess mutant-specific oncogenic dependencies in vivo. SIGNIFICANCE: Current JAK inhibitors to treat myeloproliferative neoplasms are ineffective at eradicating mutant cells. We developed an endogenously expressed Jak2V617F dual-recombinase knock-in/knock-out model to investigate Jak2V617F oncogenic reversion in vivo. Jak2V617F deletion abrogates MPN features and depletes disease-sustaining MPN stem cells, suggesting improved Jak2V617F targeting offers the potential for greater therapeutic efficacy. See related commentary by Celik and Challen, p. 701. This article is featured in Selected Articles from This Issue, p. 695.

Sequence-Dependent Conformational Transitions of Disordered Proteins During Condensation.

Intrinsically disordered proteins (IDPs) can form biomolecular condensates through phase separation. It is recognized that the conformation of IDPs in the dense and dilute phases as well as at the interfaces of condensates can critically impact the resulting properties associated with their functionality. However, a comprehensive understanding of the conformational transitions of IDPs during condensation remains elusive. In this study, we employ a coarse-grained polyampholyte model, comprising an equal number of oppositely charged residues-glutamic acid and lysine-whereby conformations and phase behavior can be readily tuned by altering the protein sequence. By manipulating the sequence patterns from perfectly alternating to block-like, we obtain chains with ideal-like conformations to semi-compact structures in the dilute phase, while in the dense phase, the chain conformation is approximately that of an ideal chain, irrespective of the protein sequence. By performing simulations at different concentrations, we find that the chains assemble from the dilute phase through small oligomeric clusters to the dense phase, accompanied by a gradual swelling of the individual chains. We further demonstrate that these findings are applicable to several naturally occurring proteins involved in the formation of biological condensates. Concurrently, we delve deeper into the chain conformations within the condensate, revealing that chains at the interface show a strong sequence dependence, but remain more collapsed than those in the bulk-like dense phase. This study addresses critical gaps in our knowledge of IDP conformations within condensates as a function of protein sequence.

Sequence-Dependent Material Properties of Biomolecular Codensates and their Relation to Dilute Phase Conformations.

Material properties of phase-separated biomolecular assemblies, enriched with disordered proteins, dictate their ability to participate in many cellular functions. Despite the significant effort dedicated to understanding how the sequence of the disordered protein drives its phase separation to form condensates, little is known about the sequence determinants of condensate material properties. Here, we computationally decipher these relationships for charged disordered proteins using model sequences comprised of glutamic acid and lysine residues as well as naturally occurring sequences of LAF1's RGG domain and DDX4's N-terminal domain. We do so by delineating how the arrangement of oppositely charged residues within these sequences influences the dynamical, rheological, and interfacial properties of the condensed phase through equilibrium and non-equilibrium molecular simulations using the hydropathy scale and Martini models. Our computations yield material properties that are quantitatively comparable with experimentally characterized condensate systems. Interestingly, we find that the material properties of both the model and natural proteins respond similarly to the segregation of charges, despite their very different sequence compositions. Condensates of the highly charge-segregated sequences exhibit slower dynamics than the uniformly charge-patterned sequences, because of their comparatively long-lived molecular contacts between oppositely charged residues. Surprisingly, the molecular interactions within the condensate are highly similar to those within a single-chain for all sequences. Consequently, the condensate material properties of charged disordered proteins are strongly correlated with their dense phase contact dynamics and their single-chain structural properties. Our findings demonstrate the potential to harness the sequence characteristics of disordered proteins for predicting and engineering the material properties of functional condensates, with insights from the dilute phase properties.

Coarse-Grained Models to Study Protein-DNA Interactions and Liquid-Liquid Phase Separation.

Recent advances in coarse-grained (CG) computational models for DNA have enabled molecular-level insights into the behavior of DNA in complex multiscale systems. However, most existing CG DNA models are not compatible with CG protein models, limiting their applications for emerging topics such as protein-nucleic acid assemblies. Here, we present a new computationally efficient CG DNA model. We first use experimental data to establish the model's ability to predict various aspects of DNA behavior, including melting thermodynamics and relevant local structural properties such as the major and minor grooves. We then employ an all-atom hydropathy scale to define nonbonded interactions between protein and DNA sites, to make our DNA model compatible with an existing CG protein model (HPS-Urry), which is extensively used to study protein phase separation, and show that our new model reasonably reproduces the experimental binding affinity for a prototypical protein-DNA system. To further demonstrate the capabilities of this new model, we simulate a full nucleosome with and without histone tails, on a microsecond time scale, generating conformational ensembles and provide molecular insights into the role of histone tails in influencing the liquid-liquid phase separation (LLPS) of HP1α proteins. We find that histone tails interact favorably with DNA, influencing the conformational ensemble of the DNA and antagonizing the contacts between HP1α and DNA, thus affecting the ability of DNA to promote LLPS of HP1α. These findings shed light on the complex molecular framework that fine-tunes the phase transition properties of heterochromatin proteins and contributes to heterochromatin regulation and function. Overall, the CG DNA model presented here is suitable to facilitate micrometer-scale studies with sub-nm resolution in many biological and engineering applications and can be used to investigate protein-DNA complexes, such as nucleosomes, or LLPS of proteins with DNA, enabling a mechanistic understanding of how molecular information may be propagated at the genome level.

The gut microbiota in adults with cystic fibrosis compared to colorectal cancer.

BACKGROUND: Gut dysbiosis is implicated in colorectal cancer (CRC) pathogenesis. Cystic fibrosis (CF) is associated with both gut dysbiosis and increased CRC risk. We therefore compared the faecal microbiota from individuals with CF to CRC and screening samples. We also assessed changes in CRC-associated taxa before and after triple CF transmembrane conductance regulator (CFTR) modulator therapy. METHODS: Bacterial DNA amplification comprising V4 16S rRNA analysis was conducted on 84 baseline and 53 matched follow-up stool samples from adults with CF. These data were compared to an existing cohort of 430 CRC and 491 control gFOBT samples from the NHS Bowel Cancer Screening Programme. Data were also compared to 26 previously identified CRC-associated taxa from a published meta-analysis. RESULTS: Faecal CF samples had a lower alpha diversity and clustered distinctly from both CRC and control samples, with no clear clinical variables explaining the variation. Compared to controls, CF samples had an increased relative abundance in 6 of the 20 enriched CRC-associated taxa and depletion of 2 of the 6 taxa which have been reported as reduced in CRC. Commencing triple modulator therapy had subtle influence on the relative abundance of CRC-associated microbiota (n = 23 paired CF samples). CONCLUSIONS: CF stool samples were clearly dysbiotic, clustering distinctly from both CRC and control samples. Several bacterial shifts in CF samples resembled those observed in CRC. Studies assessing the impact of dietary or other interventions and the longer-term use of CFTR modulators on reducing this potentially pro-oncogenic milieu are needed.

A complex network of interdomain interactions underlies the conformational ensemble of monomeric TDP-43 and modulates its phase behavior.

TAR DNA-binding protein 43 (TDP-43) is a multidomain protein involved in the regulation of RNA metabolism, and its aggregates have been observed in neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). Numerous studies indicate TDP-43 can undergo liquid-liquid phase separation (LLPS) in vitro and is a component of biological condensates. Homo-oligomerization via the folded N-terminal domain (aa:1-77) and the conserved helical region (aa:319-341) of the disordered, C-terminal domain is found to be an important driver of TDP-43 phase separation. However, a comprehensive molecular view of TDP-43 phase separation, particularly regarding the nature of heterodomain interactions, is lacking due to the challenges associated with its stability and purification. Here, we utilize all-atom and coarse-grained (CG) molecular dynamics (MD) simulations to uncover the network of interdomain interactions implicated in TDP-43 phase separation. All-atom simulations uncovered the presence of transient, interdomain interactions involving flexible linkers, RNA-recognition motif (RRM) domains and a charged segment of disordered C-terminal domain (CTD). CG simulations indicate these inter-domain interactions which affect the conformational landscape of TDP-43 in the dilute phase are also prevalent in the condensed phase. Finally, sequence and surface charge distribution analysis coupled with all-atom simulations (at high salt) confirmed that the transient interdomain contacts are predominantly electrostatic in nature. Overall, our findings from multiscale simulations lead to a greater appreciation of the complex interaction network underlying the structural landscape and phase separation of TDP-43.

Metabolic dependencies of metastasis-initiating cells in female breast cancer.

Understanding the mechanisms that enable cancer cells to metastasize is essential in preventing cancer progression. Here we examine the metabolic adaptations of metastasis-initiating cells (MICs) in female breast cancer and how those shape their metastatic phenotype. We find that endogenous MICs depend on the oxidative tricarboxylic acid cycle and fatty acid usage. Sorting tumor cells based upon solely mitochondrial membrane potential or lipid storage is sufficient at identifying MICs. We further identify that mitochondrially-generated citrate is exported to the cytoplasm to yield acetyl-CoA, and this is crucial to maintaining heightened levels of H3K27ac in MICs. Blocking acetyl-CoA generating pathways or H3K27ac-specific epigenetic writers and readers reduces expression of epithelial-to-mesenchymal related genes, MIC frequency, and metastatic potential. Exogenous supplementation of a short chain carboxylic acid, acetate, increases MIC frequency and metastasis. In patient cohorts, we observe that higher expression of oxidative phosphorylation related genes is associated with reduced distant relapse-free survival. These data demonstrate that MICs specifically and precisely alter their metabolism to efficiently colonize distant organs.

How a disordered linker in the Polycomb protein Polyhomeotic tunes phase separation and oligomerization.

The Polycomb Group (PcG) complex PRC1 represses transcription, forms condensates in cells, and modifies chromatin architecture. These processes are connected through the essential, polymerizing Sterile Alpha Motif (SAM) present in the PRC1 subunit Polyhomeotic (Ph). In vitro, Ph SAM drives formation of short oligomers and phase separation with DNA or chromatin in the context of a Ph truncation ("mini-Ph"). Oligomer length is controlled by the long disordered linker (L) that connects the SAM to the rest of Ph--replacing Drosophila PhL with the evolutionarily diverged human PHC3L strongly increases oligomerization. How the linker controls SAM polymerization, and how polymerization and the linker affect condensate formation are not know. We analyzed PhL and PHC3L using biochemical assays and molecular dynamics (MD) simulations. PHC3L promotes mini-Ph phase separation and makes it relatively independent of DNA. In MD simulations, basic amino acids in PHC3L form contacts with acidic amino acids in the SAM. Engineering the SAM to make analogous charge-based contacts with PhL increased polymerization and phase separation, partially recapitulating the effects of the PHC3L. Ph to PHC3 linker swaps and SAM surface mutations alter Ph condensate formation in cells, and Ph function in Drosophila imaginal discs. Thus, SAM-driven phase separation and polymerization are conserved between flies and mammals, but the underlying mechanisms have diverged through changes to the disordered linker.

Stimuli-responsive pressure-strain sensor-based conductive hydrogel for alleviated non-alcoholic fatty liver disease by scavenging reactive oxygen species in adipose tissue.

A visible light- and reactive oxygen species (ROS)-responsive pressure/strain sensor based on carbon dot (CD)-loaded conductive hydrogel was developed for detecting high-fat diet (HFD) and preventing the risk of non-alcoholic fatty liver disease. The designed nanoparticle consisted of a diselenide polymer dot (dsPD) loaded with a visible light-responsive CD to form dsPD@CD (DSCD). The influence of visible light irradiation and ROS on DSCD facilitated the electron transport, enhancing the conductivity of DSCD-embedded hydrogel (DSCD hydrogel) from 1.3 to 35.9 mS/m. Alternatively, the tensile modulus of the DSCD hydrogel enhanced to 223 % after light-induced ROS treatment, which simultaneously impacted the capacitive response (120 %). The hydrogel implantation into inguinal white adipose tissue of HFD mice showed 82 % higher conductivity and 83 % enhanced pressure sensing response to HFD-generated high ROS levels compared with the normal diet-fed mice. Additionally, the ROS scavenging activity of DSCD hydrogel was confirmed by the downregulation of ROS-responsive genes, such as Sod2, Nrf2, and catalase (Cat) in murine primary hepatocytes isolated from fatty liver-induced mice. In addition, in vivo animal studies also confirmed the suppression of hepatic lipogenesis, as shown by decreased Pparγ and Fasn expression and hypertrophy of adipocytes in HFD mice. The distinguishable real-time wireless resistance response observed with pressure sensing indicates the potential application of the device for monitoring the risk of non-alcoholic fatty liver disease. STATEMENT OF SIGNIFICANCE: A visible-light-induced ROS-responsive carbon dot-loaded conductive hydrogel was developed for the detection of HFD-induced alterations in ROS levels by evaluating the conductivity and electrochemical responses with applied pressure/strain. The implanted hydrogel facilitates the recovery of the inflated adipocytes induced by NAFLD, which reduces fat accumulation in the liver, preventing the risk of NAFLD. Real-time detection based on the resistance response during local compression of the hydrogel is possibly performed utilizing a wireless sensing device, demonstrating the ease of NAFLD monitoring.

Soft implantable printed bioelectronic system for wireless continuous monitoring of restenosis.

Atherosclerosis is a prominent cause of coronary artery disease and broader cardiovascular diseases, the leading cause of death worldwide. Angioplasty and stenting is a common treatment, but in-stent restenosis, where the artery re-narrows, is a frequent complication. Restenosis is detected through invasive procedures and is not currently monitored frequently for patients. Here, we report an implantable vascular bioelectronic device using a newly developed miniaturized strain sensor via microneedle printing methods. A capillary-based printing system achieves high-resolution patterning of a soft, capacitive strain sensor. Ink and printing parameters are evaluated to create a fully printed sensor, while sensor design and sensing mechanism are studied to enhance sensitivity and minimize sensor size. The sensor is integrated with a wireless vascular stent, offering a biocompatible, battery-free, wireless monitoring system compatible with conventional catheterization procedures. The vascular sensing system is demonstrated in an artery model for monitoring restenosis progression. Collectively, the artery implantable bioelectronic system shows the potential for wireless, real-time monitoring of various cardiovascular diseases and stent-integrated sensing/treatments.

Dynamics and Power Balance of Near Unity Target Gain Inertial Confinement Fusion Implosions.

The change in the power balance, temporal dynamics, emission weighted size, temperature, mass, and areal density of inertially confined fusion plasmas have been quantified for experiments that reach target gains up to 0.72. It is observed that as the target gain rises, increased rates of self-heating initially overcome expansion power losses. This leads to reacting plasmas that reach peak fusion production at later times with increased size, temperature, mass and with lower emission weighted areal densities. Analytic models are consistent with the observations and inferences for how these quantities evolve as the rate of fusion self-heating, fusion yield, and target gain increase. At peak fusion production, it is found that as temperatures and target gains rise, the expansion power loss increases to a near constant ratio of the fusion self-heating power. This is consistent with models that indicate that the expansion losses dominate the dynamics in this regime.

Generative Adversarial Network-Enhanced Ultra-Low-Dose [18F]-PI-2620 τ PET/MRI in Aging and Neurodegenerative Populations.

BACKGROUND AND PURPOSE: With the utility of hybrid τ PET/MR imaging in the screening, diagnosis, and follow-up of individuals with neurodegenerative diseases, we investigated whether deep learning techniques can be used in enhancing ultra-low-dose [18F]-PI-2620 τ PET/MR images to produce diagnostic-quality images. MATERIALS AND METHODS: Forty-four healthy aging participants and patients with neurodegenerative diseases were recruited for this study, and [18F]-PI-2620 τ PET/MR data were simultaneously acquired. A generative adversarial network was trained to enhance ultra-low-dose τ images, which were reconstructed from a random sampling of 1/20 (approximately 5% of original count level) of the original full-dose data. MR images were also used as additional input channels. Region-based analyses as well as a reader study were conducted to assess the image quality of the enhanced images compared with their full-dose counterparts. RESULTS: The enhanced ultra-low-dose τ images showed apparent noise reduction compared with the ultra-low-dose images. The regional standard uptake value ratios showed that while, in general, there is an underestimation for both image types, especially in regions with higher uptake, when focusing on the healthy-but-amyloid-positive population (with relatively lower τ uptake), this bias was reduced in the enhanced ultra-low-dose images. The radiotracer uptake patterns in the enhanced images were read accurately compared with their full-dose counterparts. CONCLUSIONS: The clinical readings of deep learning-enhanced ultra-low-dose τ PET images were consistent with those performed with full-dose imaging, suggesting the possibility of reducing the dose and enabling more frequent examinations for dementia monitoring.