Image Captioning via Dynamic Path Customization.

This article explores a novel dynamic network for vision and language (V&L) tasks, where the inferring structure is customized on the fly for different inputs. Most previous state-of-the-art (SOTA) approaches are static and handcrafted networks, which not only heavily rely on expert knowledge but also ignore the semantic diversity of input samples, therefore resulting in suboptimal performance. To address these issues, we propose a novel Dynamic Transformer Network (DTNet) for image captioning, which dynamically assigns customized paths to different samples, leading to discriminative yet accurate captions. Specifically, to build a rich routing space and improve routing efficiency, we introduce five types of basic cells and group them into two separate routing spaces according to their operating domains, i.e., spatial and channel. Then, we design a Spatial-Channel Joint Router (SCJR), which endows the model with the capability of path customization based on both spatial and channel information of the input sample. To validate the effectiveness of our proposed DTNet, we conduct extensive experiments on the MS-COCO dataset and achieve new SOTA performance on both the Karpathy split and the online test server. The source code is publicly available at https://github.com/xmu-xiaoma666/DTNet.

Characterization of Urinary N-Acetyltaurine as a Biomarker of Hyperacetatemia in Mice.

Acetate is an important metabolite in metabolic fluxes. Its presence in biological entities originates from both exogenous inputs and endogenous metabolism. Because the change in blood acetate level has been associated with both beneficial and adverse health outcomes, blood acetate analysis has been used to monitor the systemic status of acetate turnover. The present study examined the use of urinary N-acetyltaurine (NAT) as a marker to reflect the hyperacetatemic status of mice from exogenous inputs and endogenous metabolism, including triacetin dosing, ethanol dosing, and streptozotocin-induced diabetes. The results showed that triacetin dosing increased serum acetate and urinary NAT but not other N-acetylated amino acids in urine. The co-occurrences of increased serum acetate and elevated urinary NAT were also observed in both ethanol dosing and streptozotocin-induced diabetes. Furthermore, the renal cortex was determined as an active site for NAT synthesis. Overall, urinary NAT behaved as an effective marker of hyperacetatemia in three experimental mouse models, warranting further investigation into its application in humans.

Chemical composition, pharmacological effects, and parasitic mechanisms of Cistanche deserticola: An update.

BACKGROUND: C. deserticola, a highly esteemed medicinal herb in China, commonly referred to as "desert ginseng", has been renowned for its unique pharmacological properties in clinical use for countless centuries. Despite its long-standing reputation, our current comprehension of its active components and pharmacological effects remains shallow and incomplete. Moreover, the unclear mechanism underlying its pharmacological actions hinders the advancement and utilization of novel drug formulations derived from C. deserticola. Furthermore, as a unique parasitic plant, the current research on its parasitic mechanisms is limited, hampering efforts to enhance both its medicinal composition and overall yields. PURPOSE: The objective of this review is to meticulously assess, condense, and evaluate the salient aspects pertaining to the chemical composition, pharmacological impacts, and parasitic mechanisms of C. deserticola. Furthermore, the aim is to furnish valuable references that can inform and guide future research endeavors and developmental activities related to C. deserticola. METHODS: This review adheres to the rigorous standards outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A thorough examination and analysis of pertinent research findings, published up to February 6, 2024, has been conducted. Databases such as PubMed, Scopus, Web of Science, Cochrane, and Science Direct were exhaustively searched using targeted keywords and operators to delve into the chemical constituents, pharmacological effects, and parasitic mechanisms exhibited by C. deserticola. RESULTS: The review comprehensively summarizes the advancements in research regarding the chemical composition, pharmacological impacts, and toxicological safety of C. deserticola. It delves into the parasitic mechanisms of C. deserticola from three distinct angles: seed germination, haustorium induction, and recognition of signal substances. Furthermore, the review pinpoints pertinent issues and offers insightful recommendations for future exploration and research pertaining to C. deserticola. CONCLUSION: In recent years, C. deserticola has garnered considerable attention due to its distinctive pharmacological properties. This comprehensive review aims to establish a scientific foundation for the development of potential novel drugs and the enhancement of both the quantity and quality of C. deserticola. It accomplishes this by meticulously analyzing and evaluating the latest research findings pertaining to its chemical composition, pharmacological impacts, and parasitic mechanisms.

Potential role of epithelial-mesenchymal transition induced by periodontal pathogens in oral cancer.

With the increasing incidence of oral cancer in the world, it has become a hotspot to explore the pathogenesis and prevention of oral cancer. It has been proved there is a strong link between periodontal pathogens and oral cancer. However, the specific molecular and cellular pathogenic mechanisms remain to be further elucidated. Emerging evidence suggests that periodontal pathogens-induced epithelial-mesenchymal transition (EMT) is closely related to the progression of oral cancer. Cells undergoing EMT showed increased motility, aggressiveness and stemness, which provide a pro-tumour environment and promote malignant metastasis of oral cancer. Plenty of studies proposed periodontal pathogens promote carcinogenesis via EMT. In the current review, we discussed the association between the development of oral cancer and periodontal pathogens, and summarized various mechanisms of EMT caused by periodontal pathogens, which are supposed to play an important role in oral cancer, to provide targets for future research in the fight against oral cancer.

Association between the presence and genotype of Helicobacter pylori and periodontitis.

Whether Helicobacter pylori (H. pylori) infection is associated with periodontitis has been contested for decades. The relationship between H. pylori genotypes and periodontitis has not been clarified either. The present study provides a novel perspective to better understand the role of H. pylori in the pathogenesis of periodontitis. A total of 53 volunteers were recruited and divided into 3 groups in this cross-sectional study, namely the periodontally healthy group (15 participants), the stage I/II periodontitis group (20 participants) and the stage III/IV periodontitis group (18 participants). DNA from the subgingival plaque of all participants was extracted and PCR was performed using specific primers for the urease C gene and cytotoxin-associated gene A (cagA)/vacuolating cytotoxin gene A (vacA) to detect the presence and genotype of H. pylori. A χ2 test and one-way ANOVA were performed on the data. There was no significant difference in sex, age or body mass index between the groups. The detection rate of H. pylori was 39.62% in the total population and increased with the deepening of probing depth and clinical attachment loss. There were significant differences in the detection rate of H. pylori among the three groups, with 13.33, 40.00 and 61.11% in the periodontally healthy, stage I/II periodontitis and stage III/IV periodontitis groups, respectively (χ2=8.760, P<0.001). The cagA-/vacAs2m2 genotype was most commonly detected in the periodontally healthy group (100%). In the periodontitis group, cagA+/vacAs1m2 was the most commonly detected genotype in the stage I/II periodontitis group (37.5%) and cagA+/vacAs1m1 in the stage III/IV periodontitis group (36.3%). The results of the present study suggest that the detection rates and genotypes of H. pylori in the subgingival plaque are associated with the status of periodontitis. cagA+/vacAs1m1 and cagA+/vacAs1m2 may be considered virulence markers of periodontitis. However, given the small sample size and lack of correlation analysis of the study, further larger scale and high-quality clinical trials are required to confirm these findings.

Point-coated long-period fiber grating for temperature measurement.

In this paper, a temperature sensor based on a point-coated long-period fiber grating (PC-LPFG) is proposed and investigated. This structure is fabricated using a thermal filling method. The point-coating approach effectively increases the coupling efficiency between the sensing unit and the surrounding medium. The polymethyl methacrylate (PMMA), with high thermal optical coefficient (TOC) and thermal expansion coefficient (TEC), improves the temperature sensitivity of the PC-LPFG. Experimental results show that the temperature sensitivities of this sensor are 2.948 nm/°C and 6.717 nm/°C in the temperature ranges of 80.4-91°C and 91-97°C, respectively. The hot point-coating method of the PC-LPFG provides a new, to the best of our knowledge, approach to combining optic fiber sensors with high polymer materials.

RhoA/ROCK1 regulates the mitochondrial dysfunction through Drp1 induced by Porphyromonas gingivalis in endothelial cells.

Porphyromonas gingivalis (P. gingivalis) is a pivotal pathogen of periodontitis. Our previous studies have confirmed that mitochondrial dysfunction in the endothelial cells caused by P. gingivalis was dependent on Drp1, which may be the mechanism of P. gingivalis causing endothelial dysfunction. Nevertheless, the signalling pathway induced the mitochondrial dysfunction remains unclear. The purpose of this study was to investigate the role of the RhoA/ROCK1 pathway in regulating mitochondrial dysfunction caused by P. gingivalis. P. gingivalis was used to infect EA.hy926 cells (endothelial cells). The expression and activation of RhoA and ROCK1 were assessed by western blotting and pull-down assay. The morphology of mitochondria was observed by mitochondrial staining and transmission electron microscopy. Mitochondrial function was measured by ATP content, mitochondrial DNA and mitochondrial permeability transition pore openness. The phosphorylation and translocation of Drp1 were evaluated using western blotting and immunofluorescence. The role of the RhoA/ROCK1 pathway in mitochondrial dysfunction was investigated using RhoA and ROCK1 inhibitors. The activation of RhoA/ROCK1 pathway and mitochondrial dysfunction were observed in P. gingivalis-infected endothelial cells. Furthermore, RhoA or ROCK1 inhibitors partly prevented mitochondrial dysfunction caused by P. gingivalis. The increased phosphorylation and mitochondrial translocation of Drp1 induced by P. gingivalis were both blocked by RhoA and ROCK1 inhibitors. In conclusion, we demonstrate that the RhoA/ROCK1 pathway was involved in mitochondrial dysfunction caused by P. gingivalis by regulating the phosphorylation and mitochondrial translocation of Drp1. Our research illuminated a possible new mechanism by which P. gingivalis promotes endothelial dysfunction.

Highly sensitive curvature fiber sensor based on an enhanced core diameter mismatch.

In this study, a curvature fiber sensor based on an enhanced core diameter mismatch is experimentally proposed and theoretically investigated. The structure is fabricated by splicing two types of step multimode and coreless fibers to excite the high-order cladding modes to improve the curvature sensitivity. Experimental results show that the highest curvature sensitivities of the structure reach -114.74 nm/m-1 in the Dip 1272 nm, -91.08 nm/m-1 in Dip 1408 nm, and -61.10 nm/m-1 in Dip 1644 nm in the measuring range of 0-0.49778 m-1. Meanwhile, the sensor's temperature and strain responses were also tested, which shows little influence on the curvature measurement. Additionally, the proposed fiber sensor exhibits features of easy fabrication, simple structure, and high mechanical strength. This study proposes a device for curvature measurement with potential use in material mechanics and optical fiber sensor design.

Regular consumption of lacto-fermented vegetables has greater effects on the gut metabolome compared with the microbiome.

The industrialisation of Western food systems has reduced the regular consumption of lacto-fermented vegetables (LFV). Consuming LFV may exert health benefits through the alteration of the gut microbiome, but the mechanisms involved remain unclear. To start understanding the possible benefits of LFV, we compared faecal microbial diversity and composition, as well as dietary habits between individuals who regularly consume LFV (n = 23) and those who do not (n = 24). We utilised microbial DNA amplicon sequencing (16S rRNA and ITS2) and untargeted metabolomics (LC-MS) to analyse stool samples. Study participants also provided three consecutive days of dietary data. Results show minor effects on microbiome composition; with the enrichment of a few microorganisms potentially associated with vegetable ferments, such as Leuconostoc mesenteroides and Rhodotorula mucilaginosa (P < 0.05), in LFV consumers. However, LFV consumption had greater effects on the faecal metabolome, with higher abundances of butyrate, acetate, and valerate (P < 0.05) and significantly greater metabolome diversity (P < 0.001). Overall, the observations of minor changes in the faecal microbiome and greater effects on the faecal metabolome from LFV consumption warrant further investigations on the health significance of LFV as regular components of the daily diet in humans.

High-sensitivity strain sensor based on a helical-core long-period fiber grating.

A highly sensitive strain sensor based on a helical-core long-period fiber grating (HC-LPFG) is proposed and experimentally investigated. The helical core is fabricated in the common single-mode fiber by using a high-frequency CO2 laser and hydrogen-oxygen flame. This helical shape core of the structure experienced the highly centralized refractive index modulation, which enhances the strain sensitivity and shortens the length of the sensing area to 2 mm. Experimental results indicate that the maximum strain sensitivity of the HC-LPFG reaches -97 pm/µÉ› within the measuring range of 0-400 µÉ›.

Identification of Independent and Shared Metabolic Responses to High-Fiber and Antibiotic Treatments in Fecal Metabolome of Grow-Finish Pigs.

Feeding high-fiber (HF) coproducts to grow-finish pigs as a cost-saving practice could compromise growth performance, while the inclusion of antibiotic growth promoters (AGPs) may improve it. The hindgut is a shared site of actions between fiber and AGPs. However, whether the metabolic interactions between them could occur in the digestive tract of pigs and then become detectable in feces have not been well-examined. In this study, wheat middling (WM), a HF coproduct, and bacitracin, a peptide antibiotic (AB), were fed to 128 grow-finish pigs for 98 days following a 2 × 2 factorial design, including antibiotic-free (AF) + low fiber (LF); AF + HF; AB + LF, and AB + HF, for growth and metabolic responses. The growth performance of the pigs was compromised by HF feedings but not by AB. A metabolomic analysis of fecal samples collected on day 28 of feeding showed that WM elicited comprehensive metabolic changes, especially in amino acids, fatty acids, and their microbial metabolites, while bacitracin caused selective metabolic changes, including in secondary bile acids. Limited metabolic interactions occurred between fiber and AB treatments. Moreover, the correlations between individual fecal metabolites and growth support the usage of fecal metabolome as a source of biomarkers for monitoring and predicting the metabolic performance of grow-finish pigs.

Knowing What to Learn: A Metric-Oriented Focal Mechanism for Image Captioning.

Despite considerable progress, image captioning still suffers from the huge difference in quality between easy and hard examples, which is left unexploited in existing methods. To address this issue, we explore the hard example mining in image captioning, and propose a simple yet effective mechanism to instruct the model to pay more attention to hard examples, thereby improving the performance in both general and complex scenarios. We first propose a novel learning strategy, termed Metric-oriented Focal Mechanism (MFM), for hard example mining in image captioning. Differing from the existing strategies for classification tasks, MFM can adopt the generative metrics of image captioning to measure the difficulties of examples, and then up-weight the rewards of hard examples during training. To make MFM applicable to different datasets without tedious parameter tuning, we further introduce an adaptive reward metric called Effective CIDEr (ECIDEr), which considers the data distribution of easy and hard examples during reward estimation. Extensive experiments are conducted on the MS COCO benchmark, and the results show that while maintaining the performance on simple examples, MFM can significantly improve the quality of captions for hard examples. The ECIDEr-based MFM is equipped on the current SOTA method, e.g., DLCT (Luo et al., 2021), which outperforms all existing methods and achieves new state-of-the-art performance on both the off-line and on- line testing, i.e., 134.3 CIDEr for the off-line testing and 136.1 for the on- line testing of MSCOCO. To validate the generalization ability of ECIDEr-based MFM, we also apply it to another dataset, namely Flickr30k, and superior performance gains can also be obtained.

Integrating Multiculturalism Into Artificial Intelligence-Assisted Programming Lessons: Examining Inter-Ethnicity Differences in Learning Expectancy, Motivation, and Effectiveness.

Given the current popularization of computer programming and the trends of informatization and digitization, colleges have actively responded by making programming lessons compulsory for students of all disciplines. However, students from different ethnic groups often have different learning responses to such lessons due to their respective cultural backgrounds, the environment in which they grew up, and their consideration for future employment. In this study, an AI-assisted programming module was developed and used to compare the differences between multi-ethnic college students in terms of their theoretical and actual learning expectancy, motivation, and effectiveness. The module conducted analysis through the deep learning network and examined the relevant processes that the students underwent during programming lessons, as well as the types of errors they had committed. Their learning motivation for and actual learning performance in programming were then examined based on the cognitive learning theory. The results of the experiment, which involved 96 multi-ethnic college students, indicated that the two groups had dissimilar theoretical performance in terms of their expectancy and motivation for learning programming. The indigenous students' main concern was whether programming would affect their families or tribes, and this concern affected and was reflected in their learning outcomes. In contrast, the learning motivation and goals of Han Chinese students were driven by the cognition of the value of programming to themselves. The research findings can contribute toward the cognition and understanding of multi-ethnic students when learning computer programming and development of the appropriate teaching methods, and serve as a reference for subsequent research on integrating multiculturalism into computer programming lessons.

Untargeted urine metabolite profiling by mass spectrometry aided by multivariate statistical analysis to predict prostate cancer treatment outcome.

Deciphering metabolomic networks has been demonstrated to provide valuable information for diagnosing and monitoring diseases. Herein, we report a technique to monitor untargeted urine metabolites to evaluate prostate cancer aggressiveness and treatment outcome. Direct chemical profiling of urine was achieved by a combined procedure of hyphenating laser diode thermal desorption with atmospheric pressure chemical ionization mass spectrometry (LDTD-APCI-MS). We describe a conceptually new approach to monitoring preoperative urinary metabolic alterations associated with prostate cancer recurrence. By evaluating mass/charge (m/z) ratios and peak intensities of ions detected by mass spectroscopy of urine samples, we revealed that intensities at m/z 313.2740 (±0.0003) and 341.3054 (±0.0006) attributable to monoacylglycerol backbone fragments from glycerides can be statistically correlated to disease progression.

Characterization of the Key Aroma Compounds of Three Kinds of Chinese Representative Black Tea and Elucidation of the Perceptual Interactions of Methyl Salicylate and Floral Odorants.

Jinjunmei (JJM), Keemun (KM), and Dianhong (DH) are the representative black teas in China, and they have always been favored by consumers. In this study, we aim to obtain the aroma characteristic information of volatile components in black tea samples through headspace solid-phase microextraction (HS-SPME), solvent-assisted flavor evaporation (SAFE), and gas chromatography-mass spectrometry combined with gas chromatography-olfactometry technology. The results showed that 70 compounds including α-methylbenzyl alcohol (isomer of ß-phenylethanol) were identified as odorants. Among them, 39 compounds such as linalool and geraniol showed a high degree of aroma contribution. Furthermore, the Feller's additive model was used to explore the perceptual interactions among the methyl salicylate and the floral compounds (10 groups): five groups of binary compounds showed masking effect after mixing, one group showed additive effect, and four groups showed synergistic effect. The ratio (R) was compared with the aroma index (n) of Steven's law, which found a high-fitness exponential relationship. The results of this study help to provide additional and new theoretical guidance for improving the aroma quality of black tea.

Integrated marine microalgae biorefineries for improved bioactive compounds: A review.

Marine microalgae offer a promising feedstock for biofuels and other valuable compounds for biorefining and carry immense potential to contribute to a clean energy and environment future. However, it is currently not economically feasible to use marine algae to produce biofuels, and the potential bioactive chemicals account for only a small market share. The production of algal biomass with multiple valuable chemicals is closely related to the algal species, cultivation conditions, culture systems, and production modes. Thus, higher requirements for screening of dominant algal strains, developing integrated technologies with the optimum culture conditions, efficient cultivation systems, and production modes to exploit algal biomass for biorefinery applications, are all needed. This review summarizes the screening of dominant microalgae, discusses the environmental conditions that may affect the growth, as well as the culture systems and production modes, and further emphasizes the valorization options of the algal biomass, which should help to offer a sustainable approach to run a profitable marine algae production system.

Eubacterium coprostanoligenes and Methanoculleus identified as potential producers of metabolites that contribute to swine manure foaming.

AIM: Swine manure foaming is a major problem, causing damage to property, livestock, and people. Here, we identified the main chemicals and microbes that contribute to foaming. METHODS AND RESULTS: Foaming and non-foaming swine manure were sampled from farms in Iowa and Illinois. Targeted and untargeted metabolomics analyses identified chemical markers that differed between foaming and non-foaming manure and between manure layers. Microbial community analysis and metagenomics were performed on a subset of samples. Foam contained significantly higher levels of total bile acids and long chain fatty acids like palmitic, stearic and oleic acid than the other manure layers. Foam layers also had significantly higher levels of ubiquinone 9 and ubiquinone 10. The slurry layer of foaming samples contained more alanine, isoleucine/leucine, diacylglycerols (DG), phosphtatidylethanolamines, and vitamin K2, while ceramide was significantly increased in the slurry layer of non-foaming samples. Eubacterium coprostanoligenes and Methanoculleus were more abundant in foaming samples, and E. coprostanoligenes was significantly correlated with levels of DG. Genes involved in diacylglycerol biosynthesis and in the biosynthesis of branched-chain hydrophobic amino acids were overrepresented in foaming samples. CONCLUSIONS: A mechanism for manure foaming is hypothesized in which proliferation of Methanoculleus leads to excessive production of methane, while production of DG by E. coprostanoligenes and hydrophobic proteins by Methanosphaera stadtmanae facilitates bubble formation and stabilization. SIGNIFICANCE AND IMPACT OF STUDY: While some chemical and biological treatments have been developed to treat swine manure foaming, its causes remain unknown. We identified key microbes and metabolites that correlate with foaming and point to possible roles of other factors like animal feed.

Design and Analysis of a Pitch Fatigue Detection System for Adaptive Baseball Learning.

Owing to the rapid development of information and communication technologies, such as the Internet of Things, artificial intelligence, and computer vision, in recent years, the concept of smart sports has been proposed. A pitch fatigue detection method that includes acquisition, analysis, quantification, aggregation, learning, and public layers for adaptive baseball learning is proposed herein. The learning determines the fatigue index of the pitcher based on the angle of the pitcher's elbow and back as the number of pitches increases. The coach uses this auxiliary information to avoid baseball injuries during baseball learning. Results show a test accuracy rate of 89.1%, indicating that the proposed method effectively provides reference information for adaptive baseball learning.

Refractometer based on fiber Mach-Zehnder interferometer composed of two micro bending cores.

In this paper, a novel refractometer based on Mach-Zehnder Interferometer (MZI) is proposed and experimentally investigated. The MZI is composed of 2 micro bending cores (MBCs), one of which excites the cladding modes and the other couples the modes back. This structure is formed by high-frequency CO2 laser polishing and oxyhydrogen flame heating. With the unique deformation method, the interaction between the fiber core and the external status gets enhanced, moreover, higher modes in the cladding are excited, which leads to a high refractive index (RI) sensitivity. Due to the high temperature of the oxyhydrogen flame, the core of CO2 polished fiber is modulated, furthermore, the cladding shape of MBC tends to be circular. Hence, relatively small modulating regions of 500 µm can form for interference. In the experiment, 2 transmission dips are chosen for RI measuring, which possesses the wavelength of 1530.4 nm and 1600.8 nm, respectively. The RI sensitivities of the 2 transmission dips are -271.7 nm/RIU and -333.8 nm/RIU with the RI range of 1.33-1.42. The temperature characteristic is also experimentally analyzed and the temperature sensitivities of which are 0.121 nm/℃ and 0.171 nm/℃ in the range of 34℃-154℃. By solving the matrix equation, the proposed sensor can be applied for simultaneous measurement of RI and temperature.

Effects of gut microbiota and fatty acid metabolism on dyslipidemia following weight-loss diets in women: Results from a randomized controlled trial.

BACKGROUND & AIMS: In our early feeding trial among overweight and obese Chinese women, both low-carbohydrate (LC) and calorie-restricted (CR) diets reduced weight and fat mass, but only the LC diet significantly improved dyslipidemia. We aimed to investigate the impacts of altered gut microbiota, fatty acid (FAs), and acylcarnitines, markers of mitochondrial function on blood lipids. METHODS: Fecal and blood samples from 48 participants at baseline and the end of a 12-week trial were used to perform metagenomics and targeted-metabolomics including erythrocyte FAs and plasma acylcarnitines, respectively. RESULTS: The two diets altered microbial structure and co-abundance gene clusters (CAGs) at different magnitudes. After a 12-week intervention, the Bacteroidetes/Firmicutes ratio increased significantly in the LC diet (P = 0.015) but not in the CR diet, which only showed an increased trend (P = 0.28). At the microbial function level, the LC group showed lower branched-chain amino acid biosynthesis and higher serine biosynthesis than the CR group. Moreover, the LC diet reduced levels of 14:0 and 16:1n-7 FAs in the de novo lipogenesis pathway, but increased 20:5n-3 compared with the CR diet. Both groups had increased plasma acylcarnitines except that the LC group had larger elevated short-chain acylcarnitines. After backward stepwise selection, a cluster of changed CAGs, FAs and acylcarnitines were found to be associated with improved lipid profile. However, changed CAGs showed higher contribution rates in elevating HDL-cholesterol (81.6%) and reducing triglycerides (89.3%) than changed FAs and acylcarnitines. CONCLUSIONS: The two weight-loss diets induced different changes of gut microbiota, plasma acylcarnitines, and erythrocyte FAs. Changes in gut microbiota rather than FA or acylcarnitine profiles showed greater contribution to improved lipid profile in these overweight and obese Chinese women. TRIAL REGISTRATION: The trial was registered at http://clinicaltrials.gov/show/NCT01358890.