Efficient and Stable CuSCN-based Perovskite Solar Cells Achieved by Interfacial Engineering with Amidinothiourea.

Cuprous thiocyanate (CuSCN) emerges as a prime candidate among inorganic hole-transport materials, particularly suitable for the fabrication of perovskite solar cells. Nonetheless, there is an Ohmic contact degradation between the perovskite and CuSCN layers. This is induced by polar solvents and undesired purities, which reduce device efficiency and operational stability. In this work, we introduce amidinothiourea (ASU) as an intermediate layer between perovskites and CuSCN to overcome the above obstacles. The characterization results confirm that ASU-modified perovskites have eliminated trap-induced defects by strong chemical bonding between -NH- and C═S from ASU and under-coordinated ions in perovskites. The interfacial engineering based on the ASU also reduces the potential barrier between the perovskite and CuSCN layers. The ASU-treated perovskite solar cells (PSC) with a gold electrode obtains an improved power conversion efficiency (PCE) from 16.36 to 18.03%. Furthermore, after being stored for 1800 h in ambient air (relative humidity (RH) = 45%), the related device without encapsulation maintains over 90% of its initial efficiency. The further combination of ASU and carbon-tape electrodes demonstrates its potential to fabricate low-cost but stable carbon-based PSCs. This work finds a universal approach for the fabrication of efficient and stable PSCs with different device structures.

LRP1 facilitates hepatic glycogenesis by improving the insulin signaling pathway in HFD-fed mice.

BACKGROUND: LDL receptor-related protein-1 (LRP1) is a cell-surface receptor that functions in diverse physiological pathways. We previously demonstrated that hepatocyte-specific LRP1 deficiency (hLRP1KO) promotes diet-induced insulin resistance and increases hepatic gluconeogenesis in mice. However, it remains unclear whether LRP1 regulates hepatic glycogenesis. METHODS: Insulin signaling, glycogenic gene expression, and glycogen content were assessed in mice and HepG2 cells. The pcDNA 3.1 plasmid and adeno-associated virus serotype 8 vector (AAV8) were used to overexpress the truncated ß-chain (ß∆) of LRP1 both in vitro and in vivo. RESULTS: On a normal chow diet, hLRP1KO mice exhibited impaired insulin signaling and decreased glycogen content. Moreover, LRP1 expression in HepG2 cells was significantly repressed by palmitate in a dose- and time-dependent manner. Both LRP1 knockdown and palmitate treatment led to reduced phosphorylation of Akt and GSK3ß, increased levels of phosphorylated glycogen synthase (GYS), and diminished glycogen synthesis in insulin-stimulated HepG2 cells, which was restored by exogenous expression of the ß∆-chain. By contrast, AAV8-mediated hepatic ß∆-chain overexpression significantly improved the insulin signaling pathway, thus activating glycogenesis and enhancing glycogen storage in the livers of high-fat diet (HFD)-fed mice. CONCLUSION: Our data revealed that LRP1, especially its ß-chain, facilitates hepatic glycogenesis by improving the insulin signaling pathway, suggesting a new therapeutic strategy for hepatic insulin resistance-related diseases.

Mudanpioside C Discovered from Paeonia suffruticosa Andr. Acts as a Protein Disulfide Isomerase Inhibitor with Antithrombotic Activities.

Paeonia suffruticosa Andr. is a well-known landscape plant worldwide and also holds significant importance in China due to its medicinal and dietary properties. Previous studies have found that Cortex Moutan (CM), the dried root bark of P. suffruticosa, showed antiplatelet and cardioprotective effects, although the underlying mechanism and active compounds remain to be revealed. In this study, protein disulfide isomerase (PDI) inhibitors in CM were identified using a ligand-fishing method combined with the UHPLC-Q-TOF-MS assay. Further, their binding sites and inhibitory activities toward PDI were validated. The antiplatelet aggregation and antithrombotic activity were investigated. The results showed that two structurally similar compounds in CM were identified as the inhibitor for PDI with IC50 at 3.22 µM and 16.73 µM; among them Mudanpioside C (MC) is the most effective PDI inhibitor. Molecular docking, site-directed mutagenesis, and MST assay unequivocally demonstrated the specific binding of MC to the b'-x domain of PDI (Kd = 3.9 µM), acting as a potent PDI inhibitor by interacting with key amino acids K263, D292, and N298 within the b'-x domain. Meanwhile, MC could dose-dependently suppress collagen-induced platelet aggregation and interfere with platelet activation, adhesion, and spreading. Administration of MC can significantly inhibit thrombosis formation without disturbing hemostasis in mice. These findings present a promising perspective on the antithrombotic properties of CM and highlight the potential application of MC as lead compounds for targeting PDI in thrombosis therapy.

4'-O-MethylbavachalconeB Targeted 14-3-3ζ Blocking the Integrin ß3 Early Outside-In Signal to Inhibit Platelet Aggregation and Thrombosis.

14-3-3ζ protein, the key target in the regulation and control of integrin ß3 outside-in signaling, is an attractive new strategy to inhibit thrombosis without affecting hemostasis. In this study, 4'-O-methylbavachalconeB (4-O-MB) in Psoraleae Fructus was identified as a 14-3-3ζ ligand with antithrombosis activity by target fishing combined with ultra-high performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UHPLC-Q-TOF-MS) analysis. The competitive inhibition analysis showed that 4-O-MB targeted 14-3-3ζ and blocked the 14-3-3ζ/integrin ß3 interaction with inhibition constant (Ki) values of 9.98 ± 0.22 µM. Molecular docking and amino acid mutation experiments confirmed that 4-O-MB specifically bound to 14-3-3ζ through LSY9 and SER28 to regulate the 14-3-3ζ/integrin ß3 interaction. Besides, 4-O-MB affected the integrin ß3 early outside-in signal by inhibiting AKT and c-Src phosphorylation. Meanwhile, 4-O-MB could inhibit ADP-, collagen-, or thrombin-induced platelet aggregation function but had no effect on platelet adhesion to collagen-coated surfaces in vivo. Administration of 4-O-MB could significantly inhibit thrombosis formation without disturbing hemostasis in mice. These findings provide new prospects for the antithrombotic effects of Psoraleae Fructus and the potential application of 4-O-MB as lead compounds in the therapy of thrombosis by targeting 14-3-3ζ.

Evaluating the representational power of pre-trained DNA language models for regulatory genomics.

The emergence of genomic language models (gLMs) offers an unsupervised approach to learn a wide diversity of cis-regulatory patterns in the non-coding genome without requiring labels of functional activity generated by wet-lab experiments. Previous evaluations have shown pre-trained gLMs can be leveraged to improve prediction performance across a broad range of regulatory genomics tasks, albeit using relatively simple benchmark datasets and baseline models. Since the gLMs in these studies were tested upon fine-tuning their weights for each downstream task, determining whether gLM representations embody a foundational understanding of cis-regulatory biology remains an open question. Here we evaluate the representational power of pre-trained gLMs to predict and interpret cell-type-specific functional genomics data that span DNA and RNA regulation. Our findings suggest that current gLMs do not offer substantial advantages over conventional machine learning approaches that use one-hot encoded sequences. This work highlights a major limitation with current gLMs, raising potential issues in conventional pre-training strategies for the non-coding genome.

Maternal Exposure to Extreme Cold Events and Risk of Congenital Heart Defects: A Large Multicenter Study in China.

Over the past decade, extreme temperature events have become more frequent and longer in duration. Previous studies on the association between extreme cold events (ECEs) and congenital heart defects (CHDs) are few and inconsistent. We conducted a national multicenter study in 1313 hospitals in 26 provinces in China and collected a total of 14 808 high CHD-risk participants from 2013 to 2021. We evaluated the ECEs experienced by each pregnant women during the embryonic period (3-8 weeks). The results indicated that ECEs experienced by pregnant women during the embryonic period were associated with the development of fetal CHD and were more strongly associated with some specific fetal CHD subtypes, such as pulmonary stenosis, pulmonary atresia, and tetralogy of Fallot. Of the CHD burden, 2.21% (95% CI: 1.43, 2.99%)-2.40% (95% CI: 1.26, 3.55%) of fetal CHD cases were attributable to ECEs during the embryonic period. Our findings emphasize the need to pay more attention to pregnant women whose embryonic period falls during the cold season to reduce cold spell detriments to newborns.

Continuous synthesis of high-entropy alloy nanoparticles by in-flight alloying of elemental metals.

High-entropy alloy (HEA) nanoparticles (NPs) exhibit unusual combinations of functional properties. However, their scalable synthesis remains a significant challenge requiring extreme fabrication conditions. Metal salts are often employed as precursors because of their low decomposition temperatures, yet contain potential impurities. Here, we propose an ultrafast (< 100 ms), one-step method that enables the continuous synthesis of HEA NPs directly from elemental metal powders via in-flight alloying. A high-temperature plasma jet ( > 5000 K) is employed for rapid heating/cooling (103 - 105 K s-1), and demonstrates the synthesis of CrFeCoNiMo HEA NPs ( ~ 50 nm) at a high rate approaching 35 g h-1 with a conversion efficiency of 42%. Our thermofluid simulation reveals that the properties of HEA NPs can be tailored by the plasma gas which affects the thermal history of NPs. The HEA NPs demonstrate an excellent light absorption of > 96% over a wide spectrum, representing great potential for photothermal conversion of solar energy at large scales. Our work shows that the thermal plasma process developed could provide a promising route towards industrial scale production of HEA NPs.

SnO2-Perovskite Interface Engineering Based on Bifacial Passivation via Multifunctional N-(2-Acetamido)-2-aminoethanesulfonic Acid toward Efficient and Stable Solar Cells.

Bifacial passivation on both electron transport materials and perovskite light-absorbing layers as a straightforward technique is used for gaining efficient and stable perovskite solar cells (PSCs). To develop this strategy, organic molecules containing multiple functional groups can maximize the effect of defect suppression. Based on this, we introduce N-(2-acetamido)-2-aminoethanesulfonic acid (ACES) at the interface between tin oxide (SnO2) and perovskite. The synergistic effect of multiple functional groups in ACES, including amino, carbonyl (C═O), and sulfonic acid (S═O) groups, promotes charge extraction of SnO2 and provides an improved energy level alignment for charge transfer. Furthermore, S═O in ACES effectively passivates the defects of uncoordinated Pb2+ in perovskite films, resulting in enhanced crystallinity and decreased nonradiative recombination at the buried interface. The power conversion efficiency (PCE) of related PSCs increases from 20.21% to 22.65% with reduced J-V hysteresis after interface modification with ACES. Notably, upon being stored at a low relative humidity of 40 ± 5% over 2000 h and high relative humidity of 80 ± 5% over 1000 h, the unencapsulated ACES-modified device retains up to 90% and 80% of their initial PCE, respectively. This study deepens defect passivation engineering on the buried interface of perovskites for realizing efficient and stable solar cells.

Performance of Artificial Intelligence in Detecting Diabetic Macular Edema From Fundus Photography and Optical Coherence Tomography Images: A Systematic Review and Meta-analysis.

BACKGROUND: Diabetic macular edema (DME) is the leading cause of vision loss in people with diabetes. Application of artificial intelligence (AI) in interpreting fundus photography (FP) and optical coherence tomography (OCT) images allows prompt detection and intervention. PURPOSE: To evaluate the performance of AI in detecting DME from FP or OCT images and identify potential factors affecting model performances. DATA SOURCES: We searched seven electronic libraries up to 12 February 2023. STUDY SELECTION: We included studies using AI to detect DME from FP or OCT images. DATA EXTRACTION: We extracted study characteristics and performance parameters. DATA SYNTHESIS: Fifty-three studies were included in the meta-analysis. FP-based algorithms of 25 studies yielded pooled area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity of 0.964, 92.6%, and 91.1%, respectively. OCT-based algorithms of 28 studies yielded pooled AUROC, sensitivity, and specificity of 0.985, 95.9%, and 97.9%, respectively. Potential factors improving model performance included deep learning techniques, larger size, and more diversity in training data sets. Models demonstrated better performance when validated internally than externally, and those trained with multiple data sets showed better results upon external validation. LIMITATIONS: Analyses were limited by unstandardized algorithm outcomes and insufficient data in patient demographics, OCT volumetric scans, and external validation. CONCLUSIONS: This meta-analysis demonstrates satisfactory performance of AI in detecting DME from FP or OCT images. External validation is warranted for future studies to evaluate model generalizability. Further investigations may estimate optimal sample size, effect of class balance, patient demographics, and additional benefits of OCT volumetric scans.

PM2.5 constituents associated with childhood obesity and larger BMI growth trajectory: A 14-year longitudinal study.

BACKGROUND: The association of specific PM2.5 chemical constituents with childhood overweight or obesity (OWOB) remain unclear. Furthermore, the long-term impacts of PM2.5 exposure on the trajectory of children's body mass index (BMI) have not been explored. METHODS: We conducted a longitudinal study among 1,450,830 Chinese children aged 6-19 years from Beijing and Zhongshan in China during 2005-2018 to examine the associations of PM2.5 and its chemical constituents with incident OWOB risk. We extracted PM2.5 mass and five main component exposure from Tracking Air Pollution in China (TAP) dataset. Cox proportional hazards models were applied to quantify exposure-response associations. We further performed principal component analysis (PCA) to handle the multi-collinearity and used quantile g-computation (QGC) approach to analyze the impacts of exposure mixtures. Additionally, we selected 125,863 children with at least 8 physical examination measurements and combined group-based trajectory models (GBTM) with multinomial logistic regression models to explore the impacts of exposure to PM2.5 mass and five constituents on BMI and BMI Z-score trajectories during 6-19 years. RESULTS: We observed each interquartile range increment in PM2.5 exposure was significantly associated with a 5.1 % increase in the risk of incident OWOB (95 % confidence Interval [CI]: 1.036-1.066). We also found black carbon, sulfate, organic matter, often linked to fossil combustion, had comparable or larger estimates of the effect (HR = 1.139-1.153) than PM2.5. Furthermore, Exposure to PM2.5 mass, sulfate, nitrate, ammonium, organic matter and black carbon was significantly associated with an increased odds of being in a larger BMI trajectory and being assigned to persistent OWOB trajectory. CONCLUSIONS: Our findings provide evidence that the constituents mainly from fossil fuel combustion may have a perceptible influence on increased OWOB risk associated with PM2.5 exposure in China. Moreover, long-term exposure to PM2.5 contributes to an increased odds of being in a lager BMI and a persistent OWOB trajectories.

A deep learning system for predicting time to progression of diabetic retinopathy.

Diabetic retinopathy (DR) is the leading cause of preventable blindness worldwide. The risk of DR progression is highly variable among different individuals, making it difficult to predict risk and personalize screening intervals. We developed and validated a deep learning system (DeepDR Plus) to predict time to DR progression within 5 years solely from fundus images. First, we used 717,308 fundus images from 179,327 participants with diabetes to pretrain the system. Subsequently, we trained and validated the system with a multiethnic dataset comprising 118,868 images from 29,868 participants with diabetes. For predicting time to DR progression, the system achieved concordance indexes of 0.754-0.846 and integrated Brier scores of 0.153-0.241 for all times up to 5 years. Furthermore, we validated the system in real-world cohorts of participants with diabetes. The integration with clinical workflow could potentially extend the mean screening interval from 12 months to 31.97 months, and the percentage of participants recommended to be screened at 1-5 years was 30.62%, 20.00%, 19.63%, 11.85% and 17.89%, respectively, while delayed detection of progression to vision-threatening DR was 0.18%. Altogether, the DeepDR Plus system could predict individualized risk and time to DR progression over 5 years, potentially allowing personalized screening intervals.

The impact of short-term exposures to ambient NO2, O3, and their combined oxidative potential on daily mortality.

It is widely recognized that air pollution exerts substantial detrimental effects in human health and the economy. The potential for harm is closely linked to the concentrations of pollutants like nitrogen dioxide (NO2) and ozone (O3), as well as their collective oxidative potential (OX). Yet, due to the challenges of directly monitoring OX as an independent factor and the influences of different substances' varying ability to contain or convey OX, uncertainties persist regarding its actual impact. To provide further evidence to the association between short-term exposures to NO2, O3, and OX and mortality, this study conducted multi-county time-series analyses with over-dispersed generalized additive models and random-effects meta-analyses to estimate the mortality data from 2014 to 2020 in Jiangsu, China. The findings reveal that short-term exposures to these pollutants are linked to increased risks of all-cause, cardiovascular, and respiratory mortality, where NO2 demonstrates 2.11% (95% confidence interval: 1.79%, 2.42%), 2.28% (1.91%, 2.66%), and 2.91% (2.13%, 3.69%) respectively per every 10 ppb increase in concentration, and the effect of O3 is 1.11% (0.98%, 1.24%), 1.39% (1.19%, 1.59%), and 1.82% (1.39%, 2.26%), and OX is 1.77% (1.58%, 1.97%), 2.19% (1.90%, 2.48%), and 2.90% (2.29%, 3.52%). Notably, women and individuals aged over 75 years exhibit higher susceptibility to these pollutants, with NO2 showing a greater impact, especially during the warm seasons. The elevated mortality rates associated with NO2, O3, and OX underscore the significance of addressing air pollution as a pressing public health issue, especially in controlling NO2 and O3 together. Further research is needed to explore the underlying mechanisms and possible influential factors of these effects.

Association of Short-Term Co-Exposure to Particulate Matter and Ozone with Mortality Risk.

A complex regional air pollution problem dominated by particulate matter (PM) and ozone (O3) needs drastic attention since the levels of O3 and PM are not decreasing in many parts of the world. Limited evidence is currently available regarding the association between co-exposure to PM and O3 and mortality. A multicounty time-series study was used to investigate the associations of short-term exposure to PM1, PM2.5, PM10, and O3 with daily mortality from different causes, which was based on data obtained from the Mortality Surveillance System managed by the Jiangsu Province Center for Disease Control and Prevention of China and analyzed via overdispersed generalized additive models with random-effects meta-analysis. We investigated the interactions of PM and O3 on daily mortality and calculated the mortality fractions attributable to PM and O3. Our results showed that PM1 is more strongly associated with daily mortality than PM2.5, PM10, and O3, and percent increases in daily all-cause nonaccidental, cardiovascular, and respiratory mortality were 1.37% (95% confidence interval (CI), 1.22-1.52%), 1.44% (95% CI, 1.25-1.63%), and 1.63% (95% CI, 1.25-2.01%), respectively, for a 10 µg/m3 increase in the 2 day average PM1 concentration. We found multiplicative and additive interactions of short-term co-exposure to PM and O3 on daily mortality. The risk of mortality was greatest among those with higher levels of exposure to both PM (especially PM1) and O3. Moreover, excess total and cardiovascular mortality due to PM1 exposure is highest in populations with higher O3 exposure levels. Our results highlight the importance of the collaborative governance of PM and O3, providing a scientific foundation for pertinent standards and regulatory interventions.

Expanding the toolset of fluorescent covalent staining of biological samples by labeling carboxylate and phosphate groups.

Recently, fluorescent covalent staining methods have been developed for visualization of anatomical structures in cells and tissues. Coupled with expansion microscopy, these stains revealed various ultrastructural details. However, the covalently stainable chemical groups have been limited to amines, carbohydrates, and thiols. Here, we developed procedures for covalently labeling tissues for carboxylate and phosphate groups, utilizing carbodiimide crosslinker chemistry. In porcine kidney tissues, the carboxylate and phosphate stain provides 1.8-4.8-fold higher signal intensity than those from the three existing stains. In cancer cells, such stain allows 2-8-fold more accurate identification of nucleoli than the amine stain. In expansion microscopy samples, such stain reveals a variety of sub-cellular structures in tissues when combined with the amine stain. Such stain also allows imaging of lipid-based structures in cultured cells. With these advantages, this new covalent staining method further expands the toolset for fluorescent visualization of histology.

Simulation Optimization of AlGaN/GaN SBD with Field Plate Structures and Recessed Anode.

This study investigated several AlGaN/GaN Schottky Barrier Diodes (SBDs) with different designs to achieve device optimization. First, the optimal electrode spacing, etching depth, and field plate size of the devices were measured using Technology Computer-Aided Design (TCAD) software by Silvaco, and analysis of the electrical behavior of the device was based on the simulation results, and several AlGaN/GaN SBD chips were designed and prepared. The experimental results revealed that the recessed anode can increase the forward current and reduce the on-resistance. An etched depth of 30 nm could obtain a turn-on voltage of 0.75 V and a forward current density of 216 mA/mm. A breakdown voltage of 1043 V and a power figure of merit (FOM) value of 572.6 MW/cm2 was obtained with a 3 µm field plate. Experiments and simulations confirmed that the recessed anode and field plate structure could increase the breakdown voltage and forward current and improve the FOM value, resulting in higher electrical performance and a wider range of application scenarios.

Ramelteon improves blood-brain barrier of focal cerebral ischemia rats to prevent post-stroke depression via upregulating occludin.

Post-stroke depression (PSD) negatively affects the prognosis of post-stroke animals. Ramelteon has neuroprotection for chronic ischemia animals, but the effect and the biological mechanism of it on PSD is still unclear. This study explored the effects of ramelteon with prophylactic administration on blood-brain barrier in rats with middle cerebral artery occlusion (MCAO) and the oxygen-glucose deprivation/reperfusion (OGD/R) bEnd.3 cells and found that ramelteon pretreatment improved the depressive-like behaviors and decreased infarct area in MCAO rats. Also, this study found ramelteon pretreatment improved viability and inhibited permeability in OGD/R cells. In addition, this study found that MCP-1, TNF-α, and IL-1 levels were raised in the MCAO rats and that occludin protein and mRNA levels were decreased in the MCAO and the OGD/R models, while the Egr-1 level was up-regulated. All of these were antagonized by ramelteon pretreatment. In addition, overexpression of Egr-1 could reverse the effect of 100 nM ramelteon pretreatment on FITC and occludin levels in OGD/R cells. In short, this study has demonstrated that the protective effect on PSD of ramelteon pretreatment on MCAO rats is related to the development of BBB permeability and that ramelteon regulates occludin to protect the BBB by inhibiting Egr-1.

Assessment of Parafoveal Diabetic Macular Ischemia on Optical Coherence Tomography Angiography Images to Predict Diabetic Retinal Disease Progression and Visual Acuity Deterioration.

Importance: The presence of diabetic macular ischemia (DMI) on optical coherence tomography angiography (OCTA) images predicts diabetic retinal disease progression and visual acuity (VA) deterioration, suggesting an OCTA-based DMI evaluation can further enhance diabetic retinopathy (DR) management. Objective: To investigate whether an automated binary DMI algorithm using OCTA images provides prognostic value on DR progression, diabetic macular edema (DME) development, and VA deterioration in a cohort of patients with diabetes. Design, Setting, and Participants: In this cohort study, DMI assessment of superficial capillary plexus and deep capillary plexus OCTA images was performed by a previously developed deep learning algorithm. The presence of DMI was defined as images exhibiting disruption of fovea avascular zone with or without additional areas of capillary loss, while absence of DMI was defined as images presented with intact fovea avascular zone outline and normal distribution of vasculature. Patients with diabetes were recruited starting in July 2015 and were followed up for at least 4 years. Cox proportional hazards models were used to evaluate the association of the presence of DMI with DR progression, DME development, and VA deterioration. Analysis took place between June and December 2022. Main Outcomes and Measures: DR progression, DME development, and VA deterioration. Results: A total of 321 eyes from 178 patients were included for analysis (85 [47.75%] female; mean [SD] age, 63.39 [11.04] years). Over a median (IQR) follow-up of 50.41 (48.16-56.48) months, 105 eyes (32.71%) had DR progression, 33 eyes (10.28%) developed DME, and 68 eyes (21.18%) had VA deterioration. Presence of superficial capillary plexus-DMI (hazard ratio [HR], 2.69; 95% CI, 1.64-4.43; P < .001) and deep capillary plexus-DMI (HR, 3.21; 95% CI, 1.94-5.30; P < .001) at baseline were significantly associated with DR progression, whereas presence of deep capillary plexus-DMI was also associated with DME development (HR, 4.60; 95% CI, 1.15-8.20; P = .003) and VA deterioration (HR, 2.12; 95% CI, 1.01-5.22; P = .04) after adjusting for age, duration of diabetes, fasting glucose, glycated hemoglobin, mean arterial blood pressure, DR severity, ganglion cell-inner plexiform layer thickness, axial length, and smoking at baseline. Conclusions and Relevance: In this study, the presence of DMI on OCTA images demonstrates prognostic value for DR progression, DME development, and VA deterioration.

EvoAug: improving generalization and interpretability of genomic deep neural networks with evolution-inspired data augmentations.

Deep neural networks (DNNs) hold promise for functional genomics prediction, but their generalization capability may be limited by the amount of available data. To address this, we propose EvoAug, a suite of evolution-inspired augmentations that enhance the training of genomic DNNs by increasing genetic variation. Random transformation of DNA sequences can potentially alter their function in unknown ways, so we employ a fine-tuning procedure using the original non-transformed data to preserve functional integrity. Our results demonstrate that EvoAug substantially improves the generalization and interpretability of established DNNs across prominent regulatory genomics prediction tasks, offering a robust solution for genomic DNNs.

Learning single-cell chromatin accessibility profiles using meta-analytic marker genes.

MOTIVATION: Single-cell assay for transposase accessible chromatin using sequencing (scATAC-seq) is a valuable resource to learn cis-regulatory elements such as cell-type specific enhancers and transcription factor binding sites. However, cell-type identification of scATAC-seq data is known to be challenging due to the heterogeneity derived from different protocols and the high dropout rate. RESULTS: In this study, we perform a systematic comparison of seven scATAC-seq datasets of mouse brain to benchmark the efficacy of neuronal cell-type annotation from gene sets. We find that redundant marker genes give a dramatic improvement for a sparse scATAC-seq annotation across the data collected from different studies. Interestingly, simple aggregation of such marker genes achieves performance comparable or higher than that of machine-learning classifiers, suggesting its potential for downstream applications. Based on our results, we reannotated all scATAC-seq data for detailed cell types using robust marker genes. Their meta scATAC-seq profiles are publicly available at https://gillisweb.cshl.edu/Meta_scATAC. Furthermore, we trained a deep neural network to predict chromatin accessibility from only DNA sequence and identified key motifs enriched for each neuronal subtype. Those predicted profiles are visualized together in our database as a valuable resource to explore cell-type specific epigenetic regulation in a sequence-dependent and -independent manner.