Search

Abstract Retrosynthesis is an important task in organic chemistry. Recently, numerous data-driven approaches have achieved promising results in this task. However, in practice, these data-driven methods might lead to sub-optimal outcomes by making predictions based on the training data distribution, a phenomenon we refer as frequency bias. For example, in template-based approaches, low-ranked predictions are typically generated by less common templates with low confidence scores which might be too low to be comparable, and it is observed that recorded reactants can be among these low-ranked predictions. In this work, we introduce RetroRanker, a ranking model built upon graph neural networks, designed to mitigate the frequency bias in predictions of existing retrosynthesis models through re-ranking. RetroRanker incorporates potential reaction changes of each set of predicted reactants in obtaining the given product to lower the rank of chemically unreasonable predictions. The predicted re-ranked results on publicly available retrosynthesis benchmarks demonstrate that we can achieve improvement on most state-of-the-art models with RetroRanker. Our preliminary studies also indicate that RetroRanker can enhance the performance of multi-step retrosynthesis.

Abstract Retrosynthesis analysis is an important task in organic chemistry with numerous industrial applications. Previously, machine learning approaches employing natural language processing techniques achieved promising results in this task by first representing reactant molecules as strings and subsequently predicting reactant molecules using text generation or machine translation models. Chemists cannot readily derive useful insights from traditional approaches that rely largely on atom-level decoding in the string representations, because human experts tend to interpret reactions by analyzing substructures that comprise a molecule. It is well-established that some substructures are stable and remain unchanged in reactions. In this paper, we developed a substructure-level decoding model, where commonly preserved portions of product molecules were automatically extracted with a fully data-driven approach. Our model achieves improvement over previously reported models, and we demonstrate that its performance can be boosted further by enhancing the accuracy of these substructures. Analyzing substructures extracted from our machine learning model can provide human experts with additional insights to assist decision-making in retrosynthesis analysis.

Farnesyl pyrophosphate synthase (FPPS) plays an important role in the synthesis of plant secondary metabolites, but its function and molecular regulation mechanism remain unclear in Pogostemon cablin. In this study, the full-length cDNA of the FPP synthase gene from P. cablin (PcFPPS) was cloned and characterized. The expressions of PcFPPS are different among different tissues (highly in P. cablin flowers). Subcellular localization analysis in protoplasts indicated that PcFPPS was located in the cytoplasm. PcFPPS functionally complemented the lethal FPPS deletion mutation in yeast CC25. Transient overexpression of PcFPPS in P. cablin leaves accelerated terpene biosynthesis, with an ~47% increase in patchouli alcohol. Heterologous overexpression of PcFPPS in tobacco plants was achieved, and it was found that the FPP enzyme activity was significantly up-regulated in transgenic tobacco by ELISA analysis. In addition, more terpenoid metabolites, including stigmasterol, phytol, and neophytadiene were detected compared with control by GC-MS analysis. Furthermore, with dual-LUC assay and yeast one-hybrid screening, we found 220 bp promoter of PcFPPS can be bound by the nuclear-localized transcription factor PcWRKY44. Overexpression of PcWRKY44 in P. cablin upregulated the expression levels of PcFPPS and patchoulol synthase gene (PcPTS), and then promote the biosynthesis of patchouli alcohol. Taken together, these results strongly suggest the PcFPPS and its binding transcription factor PcWRKY44 play an essential role in regulating the biosynthesis of patchouli alcohol.

Abstract Background Changes of soil microbial communities are one of the main factors of continuous cropping problem. Andrographis paniculata has been reported to have replant problem in cultivation. However, little is known about the variations of rhizosphere soil microbial communities of A. paniculata under a continuous cropping system. Here, Illumina MiSeq was used to investigate the shifts of rhizospheric bacterial and fungal communities after continuous cropping of A. paniculata. Results The bacterial diversity increased whereas the fungal diversity decreased in rhizosphere soil after consecutive A. paniculata monoculture; and the soil microbial community structure differed between newly plant soil and continuous cropped soil. Taxonomic analyses further revealed that the bacterial phyla Proteobacteria, Acidobacteria and Bacteroidetes and the fungal phyla Zygomycota, Ascomycota and Cercozoa were the dominant phyla across all soil samples. The relative abundance of phyla Acidobacteria and Zygomycota were significantly increased after continuous cropping. Additionally, the most abundant bacterial genus Pseudolabrys significantly decreased, while the predominant fungal genus Mortierella increased considerably in abundance after continuous cropping. Conclusions Our results revealed the changes on diversity and composition of bacterial and fungal communities in rhizospheric soil under continuous cropping of A. paniculata. These data contributed to the understanding of soil micro-ecological environments in the rhizosphere of A. paniculata.

Abstract Background Pogostemon cablin (Blanco) Benth. (Patchouli) is an important aromatic and medicinal plant and widely used in traditional Chinese medicine as well as in the perfume industry. Patchoulol is the primary bioactive component in P. cablin, its biosynthesis has attracted widespread interests. Previous studies have surveyed the putative genes involved in patchoulol biosynthesis using next-generation sequencing method; however, technical limitations generated by short-read sequencing restrict the yield of full-length genes. Additionally, little is known about the expression pattern of genes especially patchoulol biosynthesis related genes in response to methyl jasmonate (MeJA). Our understanding of patchoulol biosynthetic pathway still remained largely incomplete to date. Results In this study, we analyzed the morphological character and volatile chemical compounds of P. cablin cv. ‘Zhanxiang’, and 39 volatile chemical components were detected in the patchouli leaf using GC-MS, most of which were sesquiterpenes. Furthermore, high-quality RNA isolated from leaves and stems of P. cablin were used to generate the first full-length transcriptome of P. cablin using PacBio isoform sequencing (Iso-Seq). In total, 9.7 Gb clean data and 82,335 full-length UniTransModels were captured. 102 transcripts were annotated as 16 encoding enzymes involved in patchouli alcohol biosynthesis. Accorded with the uptrend of patchoulol content, the vast majority of genes related to the patchoulol biosynthesis were up-regulated after MeJA treatment, indicating that MeJA led to an increasing synthesis of patchoulol through activating the expression level of genes involved in biosynthesis pathway of patchoulol. Moreover, expression pattern analysis also revealed that transcription factors participated in JA regulation of patchoulol biosynthesis were differentially expressed. Conclusions The current study comprehensively reported the morphological specificity, volatile chemical compositions and transcriptome characterization of the Chinese-cultivated P. cablin cv. ‘Zhanxiang’, these results contribute to our better understanding of the physiological and molecular features of patchouli, especially the molecular mechanism of biosynthesis of patchoulol. Our full-length transcriptome data also provides a valuable genetic resource for further studies in patchouli.

China possesses abundant species resources in genus Lithocarpus Blume, commonly known as "acorns." These nuts are traditional foods in China, holding significant potential for local specialty food and rural economic development. To explore their edible value, we compared the nutritional components and metabolites between the fruits of Lithocarpus corneus (YD) and Lithocarpus pachylepis (HL), which are cultivated widely in south of China. Non‐targeted metabolomics analysis of YD and HL was performed using ultra‐performance liquid chromatography tandem mass spectrometry technology. A total of eight classes of differential metabolites (DAMs) were detected: alkaloids, amino acids and peptides, carbohydrates, fatty acids, polyketides, shikimates and phenylpropanoids, terpenoids, and other metabolites. In the positive ion mode, 225 DAMs were identified, whereas in the negative ion mode, 116 DAMs were identified. KEGG pathway enrichment analysis showed that the differentially abundant metabolites were generally enriched in three pathways: aminoacyl‐tRNA biosynthesis, alanine, aspartate, and glutamate metabolism, and ABC transporters. The nutritional content of YD has higher moisture, fructose, and total flavonoid levels, but lower starch, protein, fat, glucose, sucrose, and total sugar content than HL. Additionally, YD contains more Ca, K, P, and Fe, but less Mg and Zn compared to HL. Analyses indicate that YD is nutrient‐rich and beneficial, making it suitable for further development. Two species of Lithocarpus contain nandrolone, suggesting a new direction for specialty food development. The unique chemical composition and rich nutritional content of Lithocarpus fruits provide a theoretical basis for their development and application as food resources. Practical Application: L. corneus and L. pachylepis are traditional Chinese nuts, rich in starch, minerals, and flavonoids, making them suitable for consumption or use as food constituents. Their fruits contain various unique natural chemical compounds, indicating high application value and development potential. [ABSTRACT FROM AUTHOR]

Retrosynthesis is an important task in organic chemistry. Recently, numerous data-driven approaches have achieved promising results in this task. However, in practice, these data-driven methods might lead to sub-optimal outcomes by making predictions based on the training data distribution, a phenomenon we refer as frequency bias. For example, in template-based approaches, low-ranked predictions are typically generated by less common templates with low confidence scores which might be too low to be comparable, and it is observed that recorded reactants can be among these low-ranked predictions. In this work, we introduce RetroRanker, a ranking model built upon graph neural networks, designed to mitigate the frequency bias in predictions of existing retrosynthesis models through re-ranking. RetroRanker incorporates potential reaction changes of each set of predicted reactants in obtaining the given product to lower the rank of chemically unreasonable predictions. The predicted re-ranked results on publicly available retrosynthesis benchmarks demonstrate that we can achieve improvement on most state-of-the-art models with RetroRanker. Our preliminary studies also indicate that RetroRanker can enhance the performance of multi-step retrosynthesis. [ABSTRACT FROM AUTHOR]

Necroptosis is a promising novel target for cervical cancer therapy. Nevertheless, differentially expressed necroptosis-related genes (NRGs) in cervical cancer and their associations with prognosis are far from fully clarified. In this study, differentially expressed NRGs (DE-NRGs) were screened out and their bio-function was elucidated. Subsequently, a prognostic scoring model based on the regression coefficients of the screened out NRGs and their corresponding mRNA expressions were constructed and validated. Finally, the survival probability of cervical cancer patients based on the constructed prognostic scoring model in 3 and 5 years was predicted and assessed. We found 17 DE-NRGs in cervical cancer tissues which were closely related to cancer progression, and most of them were significantly highly expressed. Furthermore, 3 NRG were confirmed as the prognostic signature genes from 17 DE-NRGs by regression analysis. Overall survival predicted through our prognostic scoring model was lower in the high-risk group than in the low-risk group (p < 0.05) in both the TCGA cohort and the external GEO44001 validation cohort. What's more, the prediction performance of our prognostic scoring models well verified by the ROC curve, and the risk score calculated could act as an independent prognostic factor for cervical cancer patients. The calibration curve and C-index (0.776) of the nomogram analysis suggested that the predictive performance of the nomogram was satisfactory. Our study identified and validated a necroptosis-related prognostic signature in cervical cancer, which could well predict the prognosis for cervical cancer patients. [ABSTRACT FROM AUTHOR]

Retrosynthesis analysis is an important task in organic chemistry with numerous industrial applications. Previously, machine learning approaches employing natural language processing techniques achieved promising results in this task by first representing reactant molecules as strings and subsequently predicting reactant molecules using text generation or machine translation models. Chemists cannot readily derive useful insights from traditional approaches that rely largely on atom-level decoding in the string representations, because human experts tend to interpret reactions by analyzing substructures that comprise a molecule. It is well-established that some substructures are stable and remain unchanged in reactions. In this paper, we developed a substructure-level decoding model, where commonly preserved portions of product molecules were automatically extracted with a fully data-driven approach. Our model achieves improvement over previously reported models, and we demonstrate that its performance can be boosted further by enhancing the accuracy of these substructures. Analyzing substructures extracted from our machine learning model can provide human experts with additional insights to assist decision-making in retrosynthesis analysis. Retrosynthesis is a critical task for organic chemistry with numerous industrial applications. Here, the authors build a machine learning model to learn the concept of substructures from a large reaction dataset to achieve chemist-like intuitions. [ABSTRACT FROM AUTHOR]

Background: Changes of soil microbial communities are one of the main factors of continuous cropping problem. Andrographis paniculata has been reported to have replant problem in cultivation. However, little is known about the variations of rhizosphere soil microbial communities of A. paniculata under a continuous cropping system. Here, Illumina MiSeq was used to investigate the shifts of rhizospheric bacterial and fungal communities after continuous cropping of A. paniculata. Results: The bacterial diversity increased whereas the fungal diversity decreased in rhizosphere soil after consecutive A. paniculata monoculture; and the soil microbial community structure differed between newly plant soil and continuous cropped soil. Taxonomic analyses further revealed that the bacterial phyla Proteobacteria, Acidobacteria and Bacteroidetes and the fungal phyla Zygomycota, Ascomycota and Cercozoa were the dominant phyla across all soil samples. The relative abundance of phyla Acidobacteria and Zygomycota were significantly increased after continuous cropping. Additionally, the most abundant bacterial genus Pseudolabrys significantly decreased, while the predominant fungal genus Mortierella increased considerably in abundance after continuous cropping. Conclusions: Our results revealed the changes on diversity and composition of bacterial and fungal communities in rhizospheric soil under continuous cropping of A. paniculata. These data contributed to the understanding of soil micro-ecological environments in the rhizosphere of A. paniculata. [ABSTRACT FROM AUTHOR]

The JASMONATE ZIM DOMAIN (JAZ) proteins act as negative regulators in the jasmonic acid (JA) signaling pathways of plants, and these proteins have been reported to play key roles in plant secondary metabolism mediated by JA. In this study, we firstly isolated one JAZ from P. cablin, PatJAZ6, which was characterized and revealed based on multiple alignments and a phylogenic tree analysis. The result of subcellular localization indicated that the PatJAZ6 protein was located in the nucleus of plant protoplasts. The expression level of PatJAZ6 was significantly induced by the methyl jasmonate (MeJA). Furthermore, by means of yeast two-hybrid screening, we identified two transcription factors that interact with the PatJAZ6, the PatMYC2b1 and PatMYC2b2. Virus-induced gene silencing (VIGS) of PatJAZ6 caused a decrease in expression abundance, resulting in a significant increase in the accumulation of patchouli alcohol. Moreover, we overexpressed PatJAZ6 in P. cablin, which down-regulated the patchoulol synthase expression, and then suppressed the biosynthesis of patchouli alcohol. The results demonstrate that PatJAZ6 probably acts as a repressor in the regulation of patchouli alcohol biosynthesis, contributed to a model proposed for the potential JA signaling pathway in P. cablin. [ABSTRACT FROM AUTHOR]

Background: Pogostemon cablin (Blanco) Benth. (Patchouli) is an important aromatic and medicinal plant and widely used in traditional Chinese medicine as well as in the perfume industry. Patchoulol is the primary bioactive component in P. cablin, its biosynthesis has attracted widespread interests. Previous studies have surveyed the putative genes involved in patchoulol biosynthesis using next-generation sequencing method; however, technical limitations generated by short-read sequencing restrict the yield of full-length genes. Additionally, little is known about the expression pattern of genes especially patchoulol biosynthesis related genes in response to methyl jasmonate (MeJA). Our understanding of patchoulol biosynthetic pathway still remained largely incomplete to date. Results: In this study, we analyzed the morphological character and volatile chemical compounds of P. cablin cv. 'Zhanxiang', and 39 volatile chemical components were detected in the patchouli leaf using GC-MS, most of which were sesquiterpenes. Furthermore, high-quality RNA isolated from leaves and stems of P. cablin were used to generate the first full-length transcriptome of P. cablin using PacBio isoform sequencing (Iso-Seq). In total, 9.7 Gb clean data and 82,335 full-length UniTransModels were captured. 102 transcripts were annotated as 16 encoding enzymes involved in patchouli alcohol biosynthesis. Accorded with the uptrend of patchoulol content, the vast majority of genes related to the patchoulol biosynthesis were up-regulated after MeJA treatment, indicating that MeJA led to an increasing synthesis of patchoulol through activating the expression level of genes involved in biosynthesis pathway of patchoulol. Moreover, expression pattern analysis also revealed that transcription factors participated in JA regulation of patchoulol biosynthesis were differentially expressed. Conclusions: The current study comprehensively reported the morphological specificity, volatile chemical compositions and transcriptome characterization of the Chinese-cultivated P. cablin cv. 'Zhanxiang', these results contribute to our better understanding of the physiological and molecular features of patchouli, especially the molecular mechanism of biosynthesis of patchoulol. Our full-length transcriptome data also provides a valuable genetic resource for further studies in patchouli. [ABSTRACT FROM AUTHOR]