Machine learning methods have found many applications in Raman spectroscopy, especially for the identification of chemical species. However, almost all of these methods require non-trivial preprocessing such as baseline correction and/or PCA as an essential step. Here we describe our unified solution for the identification of chemical species in which a convolutional neural network is trained to automatically identify substances according to their Raman spectrum without the need of ad-hoc preprocessing steps. We evaluated our approach using the RRUFF spectral database, comprising mineral sample data. Superior classification performance is demonstrated compared with other frequently used machine learning algorithms including the popular support vector machine.

Despite the high economic and ecological importance of forests, our knowledge of the genomic evolution of trees under salt stress remains very limited. Here we report the genome sequence of the desert poplar, Populus euphratica, which exhibits high tolerance to salt stress. Its genome is very similar and collinear to that of the closely related mesophytic congener, P. trichocarpa. However, we find that several gene families likely to be involved in tolerance to salt stress contain significantly more gene copies within the P. euphratica lineage. Furthermore, genes showing evidence of positive selection are significantly enriched in functional categories related to salt stress. Some of these genes, and others within the same categories, are significantly upregulated under salt stress relative to their expression in another salt-sensitive poplar. Our results provide an important background for understanding tree adaptation to salt stress and facilitating the genetic improvement of cultivated poplars for saline soils.

BackgroundThe mechanism of high-altitude adaptation has been studied in certain mammals. However, in avian species like the ground tit Pseudopodoces humilis, the adaptation mechanism remains unclear. The phylogeny of the ground tit is also controversial.ResultsUsing next generation sequencing technology, we generated and assembled a draft genome sequence of the ground tit. The assembly contained 1.04 Gb of sequence that covered 95.4% of the whole genome and had higher N50 values, at the level of both scaffolds and contigs, than other sequenced avian genomes. About 1.7 million SNPs were detected, 16,998 protein-coding genes were predicted and 7% of the genome was identified as repeat sequences. Comparisons between the ground tit genome and other avian genomes revealed a conserved genome structure and confirmed the phylogeny of ground tit as not belonging to the Corvidae family. Gene family expansion and positively selected gene analysis revealed genes that were related to cardiac function. Our findings contribute to our understanding of the adaptation of this species to extreme environmental living conditions.ConclusionsOur data and analysis contribute to the study of avian evolutionary history and provide new insights into the adaptation mechanisms to extreme conditions in animals.

ABSTRACT

Nano-Fe3O4 accelerated electromethanogenesis on an hour-long timescale by coupling syntrophic acetate oxidation and direct interspecies electron transfer in wetland soil.

Convolutional neural networks (CNN) have been shown to provide a good solution for classification problems that utilize data obtained from vibrational spectroscopy. Moreover, CNNs are capable of identification from noisy spectra without the need for additional preprocessing. However, their application in practical spectroscopy is limited due to two shortcomings. The e ectiveness of the classification using CNNs drops rapidly when only a small number of spectra per substance are available for training (which is a typical situation in real applications). Additionally, to accommodate new, previously unseen substance classes, the network must be retrained which is computationally intensive. Here we address these issues by reformulating a multi-class classification problem with a large number of classes, but a small number of samples per class, to a binary classification problem with su icient data available for representation learning. Namely, we define the learning task as identifying pairs of inputs as belonging to the same or di erent classes. We achieve this using a Siamese convolutional neural network. A novel sampling strategy is proposed to address the imbalance problem in training the Siamese Network. The trained network can e ectively classify samples of unseen substance classes using just a single reference sample (termed as one-shot learning in the machine learning community). Our results demonstrate be er accuracy than other practical systems to date, while allowing e ortless updates of the system's database with novel substance classes.

No abstract

Urea oxidation reaction (UOR) has been proposed to replace the formidable oxygen evolution reaction (OER) to reduce the energy consumption for producing hydrogen from electrolysis of water owing to its much lower thermodynamic oxidation potential compared to that of the OER. Therefore, exploring a highly efficient and stable hydrogen evolution and urea electrooxidation bifunctional catalyst is the key to achieve economical and efficient hydrogen production. In this paper, we report a heterostructured sulfide/phosphide catalyst (Ni 3 S 2 −Ni 3 P/ NF) synthesized via one-step thermal treatment of Ni(OH) 2 /NF, which allows the simultaneous occurrence of phosphorization and sulfuration. The obtained Ni 3 S 2 −Ni 3 P/NF catalyst shows a sheet structure with an average sheet thickness of ∼100 nm, and this sheet is composed of interconnected Ni 3 S 2 and Ni 3 P nanoparticles (∼20 nm), between which there are a large number of accessible interfaces of Ni 3 S 2 −Ni 3 P. Thus, the Ni 3 S 2 −Ni 3 P/NF exhibits superior performance for both UOR and hydrogen evolution reaction (HER). For the overall urea−water electrolysis, to achieve current densities of 10 and 100 mA cm −2 , cell voltage of only 1.43 and 1.65 V is required using this catalyst as both the anode and the cathode. Moreover, this catalyst also maintains fairly excellent stability after a long-term testing, indicating its potential for efficient and energy-saving hydrogen production. The theoretical calculation results show that the Ni atoms at the interface are the most efficient catalytically active site for the HER, and the free energy of hydrogen adsorption is closest to thermal neutrality, which is only 0.16 eV. A self-driven electron transfer at the interface, making the Ni 3 S 2 sides become electron donating while Ni 3 P sides become electron withdrawing, may be the reason for the enhancement of the UOR activity. Therefore, this work shows an easy treatment for enhancing the catalytic activity of Ni-based materials to achieve high-efficiency urea−water electrolysis.