Three Copies of Four Interferon Receptor Genes Underlie a Mild Type I Interferonopathy in Down Syndrome.

Down syndrome (DS) is characterized by the occurrence of three copies of human chromosome 21 (HSA21). HSA21 contains a cluster of four interferon receptor (IFN-R) genes: IFNAR1, IFNAR2, IFNGR2, and IL10RB. DS patients often develop mucocutaneous infections and autoimmune diseases, mimicking patients with heterozygous gain-of-function (GOF) STAT1 mutations, which enhance cellular responses to three types of interferon (IFN). A gene dosage effect at these four loci may contribute to the infectious and autoimmune manifestations observed in individuals with DS. We report high levels of IFN-αR1, IFN-αR2, and IFN-γR2 expression on the surface of monocytes and EBV-transformed-B (EBV-B) cells from studying 45 DS patients. Total and phosphorylated STAT1 (STAT1 and pSTAT1) levels were constitutively high in unstimulated and IFN-α- and IFN-γ-stimulated monocytes from DS patients but lower than those in patients with GOF STAT1 mutations. Following stimulation with IFN-α or -γ, but not with IL-6 or IL-21, pSTAT1 and IFN-γ activation factor (GAF) DNA-binding activities were significantly higher in the EBV-B cells of DS patients than in controls. These responses resemble the dysregulated responses observed in patients with STAT1 GOF mutations. Concentrations of plasma type I IFNs were high in 12% of the DS patients tested (1.8% in the healthy controls). Levels of type I IFNs, IFN-Rs, and STAT1 were similar in DS patients with and without recurrent skin infections. We performed a genome-wide transcriptomic analysis based on principal component analysis and interferon modules on circulating monocytes. We found that DS monocytes had levels of both IFN-α- and IFN-γ-inducible ISGs intermediate to those of monocytes from healthy controls and from patients with GOF STAT1 mutations. Unlike patients with GOF STAT1 mutations, patients with DS had normal circulating Th17 counts and a high proportion of terminally differentiated CD8+ T cells with low levels of STAT1 expression. We conclude a mild interferonopathy in Down syndrome leads to an incomplete penetrance at both cellular and clinical level, which is not correlate with recurrent skin bacterial or fungal infections. The constitutive upregulation of type I and type II IFN-R, at least in monocytes of DS patients, may contribute to the autoimmune diseases observed in these individuals.

Deep dermatophytosis and inherited CARD9 deficiency.

BACKGROUND: Deep dermatophytosis is a severe and sometimes life-threatening fungal infection caused by dermatophytes. It is characterized by extensive dermal and subcutaneous tissue invasion and by frequent dissemination to the lymph nodes and, occasionally, the central nervous system. The condition is different from common superficial dermatophyte infection and has been reported in patients with no known immunodeficiency. Patients are mostly from North African, consanguineous, multiplex families, which strongly suggests a mendelian genetic cause. METHODS: We studied the clinical features of deep dermatophytosis in 17 patients with no known immunodeficiency from eight unrelated Tunisian, Algerian, and Moroccan families. Because CARD9 (caspase recruitment domain-containing protein 9) deficiency has been reported in an Iranian family with invasive fungal infections, we also sequenced CARD9 in the patients. RESULTS: Four patients died, at 28, 29, 37, and 39 years of age, with clinically active deep dermatophytosis. No other severe infections, fungal or otherwise, were reported in the surviving patients, who ranged in age from 37 to 75 years. The 15 Algerian and Tunisian patients, from seven unrelated families, had a homozygous Q289X CARD9 allele, due to a founder effect. The 2 Moroccan siblings were homozygous for the R101C CARD9 allele. Both alleles are rare deleterious variants. The familial segregation of these alleles was consistent with autosomal recessive inheritance and complete clinical penetrance. CONCLUSIONS: All the patients with deep dermatophytosis had autosomal recessive CARD9 deficiency. Deep dermatophytosis appears to be an important clinical manifestation of CARD9 deficiency. (Funded by Agence Nationale pour la Recherche and others.).

Fabrication of superhydrophilic-underwater superoleophobic inorganic anti-corrosive membranes for high-efficiency oil/water separation.

The issue of oil/water separation has recently become a global concern due to the frequency of oil spills and the increase in industrial waste water. Thus, membrane-based materials with unique wettability are desired to separate both of these from a mixture. Nevertheless, the fabrication of energy efficient and stable membranes appropriate for the separation process remains challenging. Herein, synergistic superhydrophilic-underwater superoleophobic inorganic membranes were inventively created by a maneuverable galvanic displacement reaction on copper mesh. The "water-loving" meshes were then used to study gravity driven oil-water separation, where a separation efficiency (the ratio of the amount of oil remaining above the membrane after the separation process to the amount of oil in original mixture) of up to 97% was achieved for various oil-water mixtures, and furthermore the wetting properties and separating performances were maintained without further attenuation after exposure to corrosive environments. Notably, the "repelling-oil" mode can switch to a superhydrophobic mode which acts as a supplementary "oil slick absorbing" material floating above the water surface and has potential in tackling oil slick clean-up issues, in comparison to the former mode which possesses better "separation ability". In addition, the original "repelling-oil" state can be reinstated with ease. The novel method involving a "one-cyclic transformation course" abandons extra chemical addition. The facile and green route presented here acts as an excellent test for the fabrication of a dual-functioning membrane with potential use in efficient oil-water separation, even in harsh environments, and off-shore oil spill cleanup.

Inherited CARD9 deficiency in otherwise healthy children and adults with Candida species-induced meningoencephalitis, colitis, or both.

BACKGROUND: Invasive infections of the central nervous system (CNS) or digestive tract caused by commensal fungi of the genus Candida are rare and life-threatening. The known risk factors include acquired and inherited immunodeficiencies, with patients often displaying a history of multiple infections. Cases of meningoencephalitis, colitis, or both caused by Candida species remain unexplained. OBJECTIVE: We studied 5 previously healthy children and adults with unexplained invasive disease of the CNS, digestive tract, or both caused by Candida species. The patients were aged 39, 7, 17, 37, and 26 years at the time of infection and were unrelated, but each was born to consanguineous parents of Turkish (2 patients), Iranian, Moroccan, or Pakistani origin. Meningoencephalitis was reported in 3 patients, meningoencephalitis associated with colitis was reported in a fourth patient, and the fifth patient had colitis only. METHODS: Inherited caspase recruitment domain family, member 9 (CARD9) deficiency was recently reported in otherwise healthy patients with other forms of severe disease caused by Candida, Trichophyton, Phialophora, and Exophiala species, including meningoencephalitis but not colitis caused by Candida and Exophiala species. Therefore we sequenced CARD9 in the 5 patients. RESULTS: All patients were found to be homozygous for rare and deleterious mutant CARD9 alleles: R70W and Q289* for the 3 patients with Candida albicans-induced meningoencephalitis, R35Q for the patient with meningoencephalitis and colitis caused by Candida glabrata, and Q295* for the patient with Candida albicans-induced colitis. Regardless of their levels of mutant CARD9 protein, the patients' monocyte-derived dendritic cells responded poorly to CARD9-dependent fungal agonists (curdlan, heat-killed C albicans, Saccharomyces cerevisiae, and Exophiala dermatitidis). CONCLUSION: Invasive infections of the CNS or digestive tract caused by Candida species in previously healthy children and even adults might be caused by inherited CARD9 deficiency.

Inherited CARD9 deficiency in 2 unrelated patients with invasive Exophiala infection.

BACKGROUND: Exophiala species are mostly responsible for skin infections. Invasive Exophiala dermatitidis disease is a rare and frequently fatal infection, with 42 cases reported. About half of these cases had no known risk factors. Similarly, invasive Exophiala spinifera disease is extremely rare, with only 3 cases reported, all in patients with no known immunodeficiency. Autosomal recessive CARD9 deficiency has recently been reported in otherwise healthy patients with severe fungal diseases caused by Candida species, dermatophytes, or Phialophora verrucosa. METHODS: We investigated an 8-year-old girl from a nonconsanguineous Angolan kindred, who was born in France and developed disseminated E. dermatitidis disease and a 26 year-old woman from an Iranian consaguineous kindred, who was living in Iran and developed disseminated E. spinifera disease. Both patients were otherwise healthy. RESULTS: We sequenced CARD9 and found both patients to be homozygous for loss-of-function mutations (R18W and E323del). The first patient had segmental uniparental disomy of chromosome 9, carrying 2 copies of the maternal CARD9 mutated allele. CONCLUSIONS: These are the first 2 patients with inherited CARD9 deficiency and invasive Exophiala disease to be described. CARD9 deficiency should thus be considered in patients with unexplained invasive Exophiala species disease, even in the absence of other infections.

Multiple Kernel Clustering With Compressed Subspace Alignment.

Multiple kernel clustering (MKC) has recently achieved remarkable progress in fusing multisource information to boost the clustering performance. However, the O(n2) memory consumption and O(n3) computational complexity prohibit these methods from being applied into median- or large-scale applications, where n denotes the number of samples. To address these issues, we carefully redesign the formulation of subspace segmentation-based MKC, which reduces the memory and computational complexity to O(n) and O(n2) , respectively. The proposed algorithm adopts a novel sampling strategy to enhance the performance and accelerate the speed of MKC. Specifically, we first mathematically model the sampling process and then learn it simultaneously during the procedure of information fusion. By this way, the generated anchor point set can better serve data reconstruction across different views, leading to improved discriminative capability of the reconstruction matrix and boosted clustering performance. Although the integrated sampling process makes the proposed algorithm less efficient than the linear complexity algorithms, the elaborate formulation makes our algorithm straightforward for parallelization. Through the acceleration of GPU and multicore techniques, our algorithm achieves superior performance against the compared state-of-the-art methods on six datasets with comparable time cost to the linear complexity algorithms.

Antibody-Targeted TNFRSF Activation for Cancer Immunotherapy: The Role of FcγRIIB Cross-Linking.

Co-stimulation signaling in various types of immune cells modulates immune responses in physiology and disease. Tumor necrosis factor receptor superfamily (TNFRSF) members such as CD40, OX40 and CD137/4-1BB are expressed on myeloid cells and/or lymphocytes, and they regulate antigen presentation and adaptive immune activities. TNFRSF agonistic antibodies have been evaluated extensively in preclinical models, and the robust antitumor immune responses and efficacy have encouraged continued clinical investigations for the last two decades. However, balancing the toxicities and efficacy of TNFRSF agonistic antibodies remains a major challenge in the clinical development. Insights into the co-stimulation signaling biology, antibody structural roles and their functionality in immuno-oncology are guiding new advancement of this field. Leveraging the interactions between antibodies and the inhibitory Fc receptor FcγRIIB to optimize co-stimulation agonistic activities dependent on FcγRIIB cross-linking selectively in tumor microenvironment represents the current frontier, which also includes cross-linking through tumor antigen binding with bispecific antibodies. In this review, we will summarize the immunological roles of TNFRSF members and current clinical studies of TNFRSF agonistic antibodies. We will also cover the contribution of different IgG structure domains to these agonistic activities, with a focus on the role of FcγRIIB in TNFRSF cross-linking and clustering bridged by agonistic antibodies. We will review and discuss several Fc-engineering approaches to optimize Fc binding ability to FcγRIIB in the context of proper Fab and the epitope, including a cross-linking antibody (xLinkAb) model and its application in developing TNFRSF agonistic antibodies with improved efficacy and safety for cancer immunotherapy.

Relationship between maternal-infant gut microbiota and infant food allergy.

The gut microbiota plays a crucial role in food allergies. We sought to identify characteristics of the maternal gut microbiota in the third trimester and the infant gut microbiota in early life and the association of these microbiotas with infant food allergy. A total of 68 healthy pregnant women and their full-term newborns were selected from a cohort of 202 mother-infant pairs; among them, 24 infants had been diagnosed with food allergy within 1 year of age, whereas 44 infants were healthy without allergic symptoms. We collected 65 maternal fecal samples before delivery and 253 infant fecal samples at five time points following birth. Fecal samples were microbiologically analyzed using 16S rRNA gene sequencing. Holdemania abundance in the maternal gut microbiota in the third trimester was significantly higher in the non-allergy group than in the food allergy group (P = 0.036). In the infant gut microbiota, Holdemania was only found in meconium samples; its abundance did not differ significantly between the two groups. The change in the abundance of Actinobacteria over time differed between the non-allergy and food allergy groups (FA, P = 0.013; NA, P = 9.8 × 10-5), and the change in the abundance of Firmicutes over time differed significantly in the non-allergy group (P = 0.023). The abundances of genera Anaerotruncus, Roseburia, Ruminococcus, and Erysipelotricaceae were significantly different between the non-allergy and food allergy groups at different time points. Our results showed that maternal carriage of Holdemania during the third trimester strongly predicted the absence of food allergies in infants; there was no correlation between the presence of food allergies and the abundance of Holdemania in the infant gut microbiota. More dynamic fluctuations in phyla Actinobacteria and Firmicutes early in life protect against food allergy. Thus, the enrichment of the infant gut microbiota early in life with short-chain fatty acid-producing bacteria may be beneficial in preventing the development of food allergies in infants.

Association Between Breastmilk Microbiota and Food Allergy in Infants.

Regulating the composition of human breastmilk has the potential to prevent allergic diseases early in life. The composition of breastmilk is complex, comprising varying levels of oligosaccharides, immunoactive molecules, vitamins, metabolites, and microbes. Although several studies have examined the relationship between different components of breastmilk and infant food allergies, few have investigated the relationship between microorganisms in breastmilk and infant food allergy. In the present study, we selected 135 healthy pregnant women and their full-term newborns from a cohort of 202 mother-infant pairs. Among them, 69 infants were exclusively breastfed until 6 mo after birth. At follow-up, 11 of the 69 infants developed a food allergy in infancy while 22 showed no signs of allergy. Thirty-three breastmilk samples were collected within 1 mo after delivery, and 123 infant fecal samples were collected at five time points following their birth. These samples were analyzed using microbial 16S rRNA gene sequencing. The abundance and evenness of the milk microbiota and the number of differential bacteria were higher in the breastmilk samples from the non-allergy group than in those from the food allergy group. The non-allergy group showed relatively high abundance of Bifidobacterium, Akkermansia, Clostridium IV, Clostridium XIVa, Veillonella, and butyrate-producing bacteria such as Fusobacterium, Lachnospiraceae incertae sedis, Roseburia, and Ruminococcus. In contrast, the abundance of Proteobacteria, Acinetobacter, and Pseudomonas in breastmilk was higher in the food allergy group. A comparison of the changes in dominant differential breastmilk microbiota in the intestinal flora of the two groups of infants over time revealed that the changes in Bifidobacterium abundance were consistent with those in the breastmilk flora. Functional pathway prediction of breastmilk microflora showed that the enhancement of the metabolic pathways of tyrosine, tryptophan, and fatty acids was significantly different between the groups. We suggest that changes in the breastmilk microbiota can influence the development of food allergies. Breastmilk contains several microbes that have protective effects against food allergies, both by influencing the colonization of intestinal microbiota and by producing butyrate. This study may provide new ideas for improving infant health through early intervention with probiotics.

S-CUDA: Self-cleansing unsupervised domain adaptation for medical image segmentation.

Medical image segmentation tasks hitherto have achieved excellent progresses with large-scale datasets, which empowers us to train potent deep convolutional neural networks (DCNNs). However, labeling such large-scale datasets is laborious and error-prone, which leads the noisy (or incorrect) labels to be an ubiquitous problem in the real-world scenarios. In addition, data collected from different sites usually exhibit significant data distribution shift (or domain shift). As a result, noisy label and domain shift become two common problems in medical imaging application scenarios, especially in medical image segmentation, which degrade the performance of deep learning models significantly. In this paper, we identify a novel problem hidden in medical image segmentation, which is unsupervised domain adaptation on noisy labeled data, and propose a novel algorithm named "Self-Cleansing Unsupervised Domain Adaptation" (S-CDUA) to address such issue. S-CUDA sets up a realistic scenario to solve the above problems simultaneously where training data (i.e., source domain) not only shows domain shift w.r.t. unsupervised test data (i.e., target domain) but also contains noisy labels. The key idea of S-CUDA is to learn noise-excluding and domain invariant knowledge from noisy supervised data, which will be applied on the highly corrupted data for label cleansing and further data-recycling, as well as on the test data with domain shift for supervised propagation. To this end, we propose a novel framework leveraging noisy-label learning and domain adaptation techniques to cleanse the noisy labels and learn from trustable clean samples, thus enabling robust adaptation and prediction on the target domain. Specifically, we train two peer adversarial networks to identify high-confidence clean data and exchange them in companions to eliminate the error accumulation problem and narrow the domain gap simultaneously. In the meantime, the high-confidence noisy data are detected and cleansed in order to reuse the contaminated training data. Therefore, our proposed method can not only cleanse the noisy labels in the training set but also take full advantage of the existing noisy data to update the parameters of the network. For evaluation, we conduct experiments on two popular datasets (REFUGE and Drishti-GS) for optic disc (OD) and optic cup (OC) segmentation, and on another public multi-vendor dataset for spinal cord gray matter (SCGM) segmentation. Experimental results show that our proposed method can cleanse noisy labels efficiently and obtain a model with better generalization performance at the same time, which outperforms previous state-of-the-art methods by large margin. Our code can be found at https://github.com/zzdxjtu/S-cuda.

Dynamic Joint Domain Adaptation Network for Motor Imagery Classification.

Electroencephalogram (EEG) has been widely used in brain computer interface (BCI) due to its convenience and reliability. The EEG-based BCI applications are majorly limited by the time-consuming calibration procedure for discriminative feature representation and classification. Existing EEG classification methods either heavily depend on the handcrafted features or require adequate annotated samples at each session for calibration. To address these issues, we propose a novel dynamic joint domain adaptation network based on adversarial learning strategy to learn domain-invariant feature representation, and thus improve EEG classification performance in the target domain by leveraging useful information from the source session. Specifically, we explore the global discriminator to align the marginal distribution across domains, and the local discriminator to reduce the conditional distribution discrepancy between sub-domains via conditioning on deep representation as well as the predicted labels from the classifier. In addition, we further investigate a dynamic adversarial factor to adaptively estimate the relative importance of alignment between the marginal and conditional distributions. To evaluate the efficacy of our method, extensive experiments are conducted on two public EEG datasets, namely, Datasets IIa and IIb of BCI Competition IV. The experimental results demonstrate that the proposed method achieves superior performance compared with the state-of-the-art methods.

Overall survival time prediction for high-grade glioma patients based on large-scale brain functional networks.

High-grade glioma (HGG) is a lethal cancer with poor outcome. Accurate preoperative overall survival (OS) time prediction for HGG patients is crucial for treatment planning. Traditional presurgical and noninvasive OS prediction studies have used radiomics features at the local lesion area based on the magnetic resonance images (MRI). However, the highly complex lesion MRI appearance may have large individual variability, which could impede accurate individualized OS prediction. In this paper, we propose a novel concept, namely brain connectomics-based OS prediction. It is based on presurgical resting-state functional MRI (rs-fMRI) and the non-local, large-scale brain functional networks where the global and systemic prognostic features rather than the local lesion appearance are used to predict OS. We propose that the connectomics features could capture tumor-induced network-level alterations that are associated with prognosis. We construct both low-order (by means of sparse representation with regional rs-fMRI signals) and high-order functional connectivity (FC) networks (characterizing more complex multi-regional relationship by synchronized dynamics FC time courses). Then, we conduct a graph-theoretic analysis on both networks for a jointly, machine-learning-based individualized OS prediction. Based on a preliminary dataset (N = 34 with bad OS, mean OS, ~400 days; N = 34 with good OS, mean OS, ~1030 days), we achieve a promising OS prediction accuracy (86.8%) on separating the individuals with bad OS from those with good OS. However, if using only conventionally derived descriptive features (e.g., age and tumor characteristics), the accuracy is low (63.2%). Our study highlights the importance of the rs-fMRI and brain functional connectomics for treatment planning.

Multi-Channel 3D Deep Feature Learning for Survival Time Prediction of Brain Tumor Patients Using Multi-Modal Neuroimages.

High-grade gliomas are the most aggressive malignant brain tumors. Accurate pre-operative prognosis for this cohort can lead to better treatment planning. Conventional survival prediction based on clinical information is subjective and could be inaccurate. Recent radiomics studies have shown better prognosis by using carefully-engineered image features from magnetic resonance images (MRI). However, feature engineering is usually time consuming, laborious and subjective. Most importantly, the engineered features cannot effectively encode other predictive but implicit information provided by multi-modal neuroimages. We propose a two-stage learning-based method to predict the overall survival (OS) time of high-grade gliomas patient. At the first stage, we adopt deep learning, a recently dominant technique of artificial intelligence, to automatically extract implicit and high-level features from multi-modal, multi-channel preoperative MRI such that the features are competent of predicting survival time. Specifically, we utilize not only contrast-enhanced T1 MRI, but also diffusion tensor imaging (DTI) and resting-state functional MRI (rs-fMRI), for computing multiple metric maps (including various diffusivity metric maps derived from DTI, and also the frequency-specific brain fluctuation amplitude maps and local functional connectivity anisotropy-related metric maps derived from rs-fMRI) from 68 high-grade glioma patients with different survival time. We propose a multi-channel architecture of 3D convolutional neural networks (CNNs) for deep learning upon those metric maps, from which high-level predictive features are extracted for each individual patch of these maps. At the second stage, those deeply learned features along with the pivotal limited demographic and tumor-related features (such as age, tumor size and histological type) are fed into a support vector machine (SVM) to generate the final prediction result (i.e., long or short overall survival time). The experimental results demonstrate that this multi-model, multi-channel deep survival prediction framework achieves an accuracy of 90.66%, outperforming all the competing methods. This study indicates highly demanded effectiveness on prognosis of deep learning technique in neuro-oncological applications for better individualized treatment planning towards precision medicine.

Exploring diagnosis and imaging biomarkers of Parkinson's disease via iterative canonical correlation analysis based feature selection.

Parkinson's disease (PD) is a neurodegenerative disorder that progressively hampers the brain functions and leads to various movement and non-motor symptoms. However, it is difficult to attain early-stage PD diagnosis based on the subjective judgment of physicians in clinical routines. Therefore, automatic and accurate diagnosis of PD is highly demanded, so that the corresponding treatment can be implemented more appropriately. In this paper, we focus on finding the most discriminative features from different brain regions in PD through T1-weighted MR images, which can help the subsequent PD diagnosis. Specifically, we proposed a novel iterative canonical correlation analysis (ICCA) feature selection method, aiming at exploiting MR images in a more comprehensive manner and fusing features of different types into a common space. To state succinctly, we first extract the feature vectors from the gray matter and the white matter tissues separately, represented as insights of two different anatomical feature spaces for the subject's brain. The ICCA feature selection method aims at iteratively finding the optimal feature subset from two sets of features that have inherent high correlation with each other. In experiments we have conducted thorough investigations on the optimal feature set extracted by our ICCA method. We also demonstrate that using the proposed feature selection method, the PD diagnosis performance is further improved, and also outperforms many state-of-the-art methods.

Multi-Label Nonlinear Matrix Completion With Transductive Multi-Task Feature Selection for Joint MGMT and IDH1 Status Prediction of Patient With High-Grade Gliomas.

The O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation and isocitrate dehydrogenase 1 (IDH1) mutation in high-grade gliomas (HGG) have proven to be the two important molecular indicators associated with better prognosis. Traditionally, the statuses of MGMT and IDH1 are obtained via surgical biopsy, which has limited their wider clinical implementation. Accurate presurgical prediction of their statuses based on preoperative multimodal neuroimaging is of great clinical value for a better treatment plan. Currently, the available data set associated with this study has several challenges, such as small sample size and complex, nonlinear (image) feature-to-(molecular) label relationship. To address these issues, we propose a novel multi-label nonlinear matrix completion (MNMC) model to jointly predict both MGMT and IDH1 statuses in a multi-task framework. Specifically, we first employ a nonlinear random Fourier feature mapping to improve the linear separability of the data, and then use transductive multi-task feature selection (performed in a nonlinearly transformed feature space) to refine the imputed soft labels, thus alleviating the overfitting problem caused by small sample size. We further design an optimization algorithm with a guaranteed convergence ability based on a block prox-linear method to solve the proposed MNMC model. Finally, by using a single-center, multimodal brain imaging and molecular pathology data set of HGG, we derive brain functional and structural connectomics features to jointly predict MGMT and IDH1 statuses. Results demonstrate that our proposed method outperforms the previously widely used single- and multi-task machine learning methods. This paper also shows the promise of utilizing brain connectomics for HGG prognosis in a non-invasive manner.

Multi-label Inductive Matrix Completion for Joint MGMT and IDH1 Status Prediction for Glioma Patients.

MGMT promoter methylation and IDH1 mutation in high-grade gliomas (HGG) have proven to be the two important molecular indicators associated with better prognosis. Traditionally, the statuses of MGMT and IDH1 are obtained via surgical biopsy, which is laborious, invasive and time-consuming. Accurate presurgical prediction of their statuses based on preoperative imaging data is of great clinical value towards better treatment plan. In this paper, we propose a novel Multi-label Inductive Matrix Completion (MIMC) model, highlighted by the online inductive learning strategy, to jointly predict both MGMT and IDH1 statuses. Our MIMC model not only uses the training subjects with possibly missing MGMT/IDH1 labels, but also leverages the unlabeled testing subjects as a supplement to the limited training dataset. More importantly, we learn inductive labels, instead of directly using transductive labels, as the prediction results for the testing subjects, to alleviate the overfitting issue in small-sample-size studies. Furthermore, we design an optimization algorithm with guaranteed convergence based on the block coordinate descent method to solve the multivariate non-smooth MIMC model. Finally, by using a precious single-center multi-modality presurgical brain imaging and genetic dataset of primary HGG, we demonstrate that our method can produce accurate prediction results, outperforming the previous widely-used single- or multi-task machine learning methods. This study shows the promise of utilizing imaging-derived brain connectome phenotypes for prognosis of HGG in a non-invasive manner.

DNMT1 cooperates with MBD4 to inhibit the expression of Glucocorticoid-induced TNFR-related protein in human T cells.

Glucocorticoid-induced TNFR-related protein (GITR) is constitutively expressed in T regulatory (Treg) cells and regulates their suppressive function. We identified two methylated CpG islands in the Gitr locus. Using a ChIP assay, we demonstrate that both DNMT1 and methyl-CpG-binding domain Protein 4 (MBD4) bind to the Gitr promoter. Moreover, knockdown of DNMT1 decreases the binding activity of MBD4. We observed much higher levels of both DNMT1 and MBD4 in human CD4+ CD25- conventional T (Tconv) cells. Moreover, co-overexpression of DNMT1 and MBD4 in Treg cells significantly inhibits GITR expression and impairs their suppressive activity. Our results reveal a novel molecular mechanism by which MBD4 inhibits GITR expression in a DNMT1-dependent manner.

Utilizing a Newly Designed Scale for Evaluating Family Support and Its Association with Exclusive Breastfeeding.

BACKGROUND: Although a woman's perception of her family members' support has long been established to be an influential factor on exclusive breastfeeding (EBF), it still has not been specified and quantified as a facilitator and guidance for practice. OBJECTIVE: To investigate in new mothers the association between EBF and maternal perception of family support with a standardized scale that classified support into nine items of behavioral or psychological support. METHOD: A cross-sectional survey was carried out among 655 new mothers to collect information on their breastfeeding behavior and their corresponding family support at a baby-friendly hospital in Beijing, China. Additionally, a nine-item standardized scale was used to explore the perceived family support for breastfeeding by new mothers. Breastfeeding behaviors were investigated using the indicators recommended by the Multiple Indicator Cluster Surveys. RESULTS: The EBF rate was 37.9%. The average score on the family perception scale reported by respondents was 28.34 ± 3.84. The new mothers who performed EBF and who predominantly breastfed perceived greater family support (29.55 ± 3.53; 29.36 ± 4.09) compared with those who performed complementary feeding or mixed feeding (26.69 ± 3.33) and those who performed artificial feeding (26.17 ± 3.14) (F = 30.296, p < 0.001). A binary logistic regression model was applied, and a stepwise regression analysis was performed with these factors; it showed that mothers with a positive perception of family support were more likely to practice EBF than those with a negative perception (adjusted odds ratio = 3.971; 95% confidence interval 2.62-6.01; p < 0.001). DISCUSSION: The EBF rate was quite low in the population investigated. Family support for breastfeeding could be evaluated by a scale, and new mothers' breastfeeding behaviors were strongly associated with their perceived family support for breastfeeding. CONCLUSION: Community healthcare providers should play a more important role in issues regarding breastfeeding among new mothers, and family support should be encouraged by health workers.

3D Deep Learning for Multi-modal Imaging-Guided Survival Time Prediction of Brain Tumor Patients.

High-grade glioma is the most aggressive and severe brain tumor that leads to death of almost 50% patients in 1-2 years. Thus, accurate prognosis for glioma patients would provide essential guidelines for their treatment planning. Conventional survival prediction generally utilizes clinical information and limited handcrafted features from magnetic resonance images (MRI), which is often time consuming, laborious and subjective. In this paper, we propose using deep learning frameworks to automatically extract features from multi-modal preoperative brain images (i.e., T1 MRI, fMRI and DTI) of high-grade glioma patients. Specifically, we adopt 3D convolutional neural networks (CNNs) and also propose a new network architecture for using multi-channel data and learning supervised features. Along with the pivotal clinical features, we finally train a support vector machine to predict if the patient has a long or short overall survival (OS) time. Experimental results demonstrate that our methods can achieve an accuracy as high as 89.9% We also find that the learned features from fMRI and DTI play more important roles in accurately predicting the OS time, which provides valuable insights into functional neuro-oncological applications.

Feature Selection Based on Iterative Canonical Correlation Analysis for Automatic Diagnosis of Parkinson's Disease.

Parkinson's disease (PD) is a major progressive neurodegenerative disorder. Accurate diagnosis of PD is crucial to control the symptoms appropriately. However, its clinical diagnosis mostly relies on the subjective judgment of physicians and the clinical symptoms that often appear late. Recent neuroimaging techniques, along with machine learning methods, provide alternative solutions for PD screening. In this paper, we propose a novel feature selection technique, based on iterative canonical correlation analysis (ICCA), to investigate the roles of different brain regions in PD through T1-weighted MR images. First of all, gray matter and white matter tissue volumes in brain regions of interest are extracted as two feature vectors. Then, a small group of significant features were selected using the iterative structure of our proposed ICCA framework from both feature vectors. Finally, the selected features are used to build a robust classifier for automatic diagnosis of PD. Experimental results show that the proposed feature selection method results in better diagnosis accuracy, compared to the baseline and state-of-the-art methods.