Durian (Durio zibethinus) is a Southeast Asian tropical plant known for its hefty, spine-covered fruit and sulfury and onion-like odor. Here we present a draft genome assembly of D. zibethinus, representing the third plant genus in the Malvales order and first in the Helicteroideae subfamily to be sequenced. Single-molecule sequencing and chromosome contact maps enabled assembly of the highly heterozygous durian genome at chromosome-scale resolution. Transcriptomic analysis showed upregulation of sulfur-, ethylene-, and lipid-related pathways in durian fruits. We observed paleopolyploidization events shared by durian and cotton and durian-specific gene expansions in MGL (methionine γ-lyase), associated with production of volatile sulfur compounds (VSCs). MGL and the ethylene-related gene ACS (aminocyclopropane-1-carboxylic acid synthase) were upregulated in fruits concomitantly with their downstream metabolites (VSCs and ethylene), suggesting a potential association between ethylene biosynthesis and methionine regeneration via the Yang cycle. The durian genome provides a resource for tropical fruit biology and agronomy.

Protein-coding mutations in clear cell renal cell carcinoma (ccRCC) have been extensively characterized, frequently involving inactivation of the von Hippel–Lindau (VHL) tumor suppressor. Roles for noncoding cis-regulatory aberrations in ccRCC tumorigenesis, however, remain unclear. Analyzing 10 primary tumor/normal pairs and 9 cell lines across 79 chromatin profiles, we observed pervasive enhancer malfunction in ccRCC, with cognate enhancer-target genes associated with tissue-specific aspects of malignancy. Superenhancer profiling identified ZNF395 as a ccRCC-specific and VHL-regulated master regulator whose depletion causes near-complete tumor elimination in vitro and in vivo. VHL loss predominantly drives enhancer/superenhancer deregulation more so than promoters, with acquisition of active enhancer marks (H3K27ac, H3K4me1) near ccRCC hallmark genes. Mechanistically, VHL loss stabilizes HIF2α–HIF1β heterodimer binding at enhancers, subsequently recruiting histone acetyltransferase p300 without overtly affecting preexisting promoter–enhancer interactions. Subtype-specific driver mutations such as VHL may thus propagate unique pathogenic dependencies in ccRCC by modulating epigenomic landscapes and cancer gene expression. Significance: Comprehensive epigenomic profiling of ccRCC establishes a compendium of somatically altered cis-regulatory elements, uncovering new potential targets including ZNF395, a ccRCC master regulator. Loss of VHL, a ccRCC signature event, causes pervasive enhancer malfunction, with binding of enhancer-centric HIF2α and recruitment of histone acetyltransferase p300 at preexisting lineage-specific promoter–enhancer complexes. Cancer Discov; 7(11); 1284–305. ©2017 AACR. See related commentary by Ricketts and Linehan, p. 1221. This article is highlighted in the In This Issue feature, p. 1201

Regulatory enhancer elements in solid tumours remain poorly characterized. Here we apply micro-scale chromatin profiling to survey the distal enhancer landscape of primary gastric adenocarcinoma (GC), a leading cause of global cancer mortality. Integrating 110 epigenomic profiles from primary GCs, normal gastric tissues and cell lines, we highlight 36,973 predicted enhancers and 3,759 predicted super-enhancers respectively. Cell-line-defined super-enhancers can be subclassified by their somatic alteration status into somatic gain, loss and unaltered categories, each displaying distinct epigenetic, transcriptional and pathway enrichments. Somatic gain super-enhancers are associated with complex chromatin interaction profiles, expression patterns correlated with patient outcome and dense co-occupancy of the transcription factors CDX2 and HNF4α. Somatic super-enhancers are also enriched in genetic risk SNPs associated with cancer predisposition. Our results reveal a genome-wide reprogramming of the GC enhancer and super-enhancer landscape during tumorigenesis, contributing to dysregulated local and regional cancer gene expression.

Integrating prior knowledge of neurophysiology into neural network architecture enhances the performance of emotion decoding. While numerous techniques emphasize learning spatial and short-term temporal patterns, there has been a limited emphasis on capturing the vital long-term contextual information associated with emotional cognitive processes. In order to address this discrepancy, we introduce a novel transformer model called emotion transformer (EmT). EmT is designed to excel in both generalized cross-subject electroencephalography (EEG) emotion classification and regression tasks. In EmT, EEG signals are transformed into a temporal graph format, creating a sequence of EEG feature graphs using a temporal graph construction (TGC) module. A novel residual multiview pyramid graph convolutional neural network (RMPG) module is then proposed to learn dynamic graph representations for each EEG feature graph within the series, and the learned representations of each graph are fused into one token. Furthermore, we design a temporal contextual transformer (TCT) module with two types of token mixers to learn the temporal contextual information. Finally, the task-specific output (TSO) module generates the desired outputs. Experiments on four publicly available datasets show that EmT achieves higher results than the baseline methods for both EEG emotion classification and regression tasks. The code is available at https://github.com/yi-ding-cs/EmT.

Carcinoma of the oral tongue (OTSCC) is the most common malignancy of the oral cavity, characterized by frequent recurrence and poor survival. The last three decades has witnessed a change in the OTSCC epidemiological profile, with increasing incidence in younger patients, females and never-smokers. Here, we sought to characterize the OTSCC genomic landscape and to determine factors that may delineate the genetic basis of this disease, inform prognosis and identify targets for therapeutic intervention. Seventy-eight cases were subjected to whole-exome (n = 18) and targeted deep sequencing (n = 60). While the most common mutation was in TP53, the OTSCC genetic landscape differed from previously described cohorts of patients with head and neck tumors: OTSCCs demonstrated frequent mutations in DST and RNF213, while alterations in CDKN2A and NOTCH1 were significantly less frequent. Despite a lack of previously reported NOTCH1 mutations, integrated analysis showed enrichments of alterations affecting Notch signaling in OTSCC. Importantly, these Notch pathway alterations were prognostic on multivariate analyses. A high proportion of OTSCCs also presented with alterations in drug targetable and chromatin remodeling genes. Patients harboring mutations in actionable pathways were more likely to succumb from recurrent disease compared with those who did not, suggesting that the former should be considered for treatment with targeted compounds in future trials. Our study defines the Asian OTSCC mutational landscape, highlighting the key role of Notch signaling in oral tongue tumorigenesis. We also observed somatic mutations in multiple therapeutically relevant genes, which may represent candidate drug targets in this highly lethal tumor type.

Previous studies have shown that each edible oil type has its own characteristic fatty acid profile; however, no method has yet been described allowing the identification of oil types simply based on this characteristic. Moreover, the fatty acid profile of a specific oil type can be mimicked by a mixture of 2 or more oil types. This has led to fraudulent oil adulteration and intentional mislabeling of edible oils threatening food safety and endangering public health. Here, we present a machine learning method to uncover fatty acid patterns discriminative for ten different plant oil types and their intra-variability. We also describe a supervised end-to-end learning method that can be generalized to oil composition of any given mixtures. Trained on a large number of simulated oil mixtures, independent test dataset validation demonstrates that the model has a 50^th percentile absolute error between 1.4–1.8% and a 90^th percentile error of 4–5.4% for any 3-way mixtures of the ten oil types. The deep learning model can also be further refined with on-line training. Because oil-producing plants have diverse geographical origins and hence slightly varying fatty acid profiles, an online-training method provides also a way to capture useful knowledge presently unavailable. Our method allows the ability to control product quality, determining the fair price of purchased oils and in-turn allowing health-conscious consumers the future of accurate labeling.