Search

Abstract: In the silkworm Bombyx mori, non-susceptibility to the Zhenjiang (China) strain of the densonucleosis virus (DNV-Z) is controlled by the recessive gene nsd-Z (non-susceptible to DNV-Z), which is located [...]

The latest progresses of experimental biology have generated a large number of data with different formats and lengths. Deep learning is an ideal tool to deal with complex datasets, but its inherent "black box" nature needs more interpretability. At the same time, traditional interpretable machine learning methods, such as linear regression or random forest, could only deal with numerical features instead of modular features often encountered in the biological field. Here, we present MultiCapsNet (https://github.com/wanglf19/MultiCapsNet), a new deep learning model built on CapsNet and scCapsNet, which possesses the merits such as easy data integration and high model interpretability. To demonstrate the ability of this model as an interpretable classifier to deal with modular inputs, we test MultiCapsNet on three datasets with different data type and application scenarios. Firstly, on the labeled variant call dataset, MultiCapsNet shows a similar classification performance with neural network model, and provides importance scores for data sources directly without an extra importance determination step required by the neural network model. The importance scores generated by these two models are highly correlated. Secondly, on single cell RNA sequence (scRNA-seq) dataset, MultiCapsNet integrates information about protein-protein interaction (PPI), and protein-DNA interaction (PDI). The classification accuracy of MultiCapsNet is comparable to the neural network and random forest model. Meanwhile, MultiCapsNet reveals how each transcription factor (TF) or PPI cluster node contributes to classification of cell type. Thirdly, we made a comparison between MultiCapsNet and SCENIC. The results show several cell type relevant TFs identified by both methods, further proving the validity and interpretability of the MultiCapsNet. [ABSTRACT FROM AUTHOR]

Abstract Background Brown planthopper (BPH), Nilaparvata lugens Stål, is one of the most destructive insect pests of rice. The molecular responses of plants to sucking insects resemble responses to pathogen infection. However, the molecular mechanism of BPH-resistance in rice remains unclear. Transcription factors (TF) are up-stream regulators of various genes that bind to specific DNA sequences, thereby controlling the transcription from DNA to mRNA. They are key regulators for transcriptional expression in biological processes, and are probably involved in the BPH-induced pathways in resistant rice varieties. Results We conducted a microarray experiment to analyze TF genes related to BPH resistance in a Sri Lankan rice cultivar, Rathu Heenati (RHT). We compared the expression profiles of TF genes in RHT with those of the susceptible rice cultivar Taichun Native 1 (TN1). We detected 2038 TF genes showing differential expression signals between the two rice varieties. Of these, 442 TF genes were probably related to BPH-induced resistance in RHT and TN1, and 229 may be related to constitutive resistance only in RHT. These genes showed a fold change (FC) of more than 2.0 (P10, there were 37 induced TF genes and 26 constitutive resistance TF genes. Of these, 13 were probably involved in BPH-induced resistance, and 8 in constitutive resistance to BPH in RHT. Conclusions We explored the molecular mechanism of resistance to BPH in rice by comparing expressions of TF genes between RHT and TN1. We speculate that the level of gene repression, especially for early TF genes, plays an important role in the defense response. The fundamental point of the resistance strategy is that plants protect themselves by reducing their metabolic level to inhibit feeding by BPH and prevent damage from water and nutrient loss. We have selected 21 TF genes related to BPH resistance for further analyses to understand the molecular responses to BPH feeding in rice.

Abstract Background Bombyx mori, the domesticated silkworm, is a well-studied model insect with great economic and scientific significance. Although more than 400 mutations have been described in silkworms, most have not been identified, especially those affecting economically-important traits. Simple sequence repeats (SSRs) are effective and economical tools for mapping traits and genetic improvement. The current SSR linkage map is of low density and contains few polymorphisms. The purpose of this work was to develop a dense and informative linkage map that would assist in the preliminary mapping and dissection of quantitative trait loci (QTL) in a variety of silkworm strains. Results Through an analysis of > 50,000 genotypes across new mapping populations, we constructed two new linkage maps covering 27 assigned chromosomes and merged the data with previously reported data sets. The integrated consensus map contains 692 unique SSR sites, improving the density from 6.3 cM in the previous map to 4.8 cM. We also developed 497 confirmed neighboring markers for corresponding low-polymorphism sites, with 244 having polymorphisms. Large-scale statistics on the SSR type were suggestive of highly efficient markers, based upon which we searched 16,462 available genomic scaffolds for SSR loci. With the newly constructed map, we mapped single-gene traits, the QTL of filaments, and a number of ribosomal protein genes. Conclusion The integrated map produced in this study is a highly efficient genetic tool for the high-throughput mapping of single genes and QTL. Compared to previous maps, the current map offers a greater number of markers and polymorphisms; thus, it may be used as a resource for marker-assisted breeding.