BERT-Kcr: prediction of lysine crotonylation sites by a transfer learning method with pre-trained BERT models.

MOTIVATION: As one of the most important post-translational modifications (PTMs), protein lysine crotonylation (Kcr) has attracted wide attention, which involves in important physiological activities, such as cell differentiation and metabolism. However, experimental methods are expensive and time-consuming for Kcr identification. Instead, computational methods can predict Kcr sites in silico with high efficiency and low cost. RESULTS: In this study, we proposed a novel predictor, BERT-Kcr, for protein Kcr sites prediction, which was developed by using a transfer learning method with pre-trained bidirectional encoder representations from transformers (BERT) models. These models were originally used for natural language processing (NLP) tasks, such as sentence classification. Here, we transferred each amino acid into a word as the input information to the pre-trained BERT model. The features encoded by BERT were extracted and then fed to a BiLSTM network to build our final model. Compared with the models built by other machine learning and deep learning classifiers, BERT-Kcr achieved the best performance with AUROC of 0.983 for 10-fold cross validation. Further evaluation on the independent test set indicates that BERT-Kcr outperforms the state-of-the-art model Deep-Kcr with an improvement of about 5% for AUROC. The results of our experiment indicate that the direct use of sequence information and advanced pre-trained models of NLP could be an effective way for identifying PTM sites of proteins. AVAILABILITY AND IMPLEMENTATION: The BERT-Kcr model is publicly available on http://zhulab.org.cn/BERT-Kcr_models/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Identification and Expression Analysis of Strigolactone Biosynthetic and Signaling Genes in Response to Salt and Alkaline Stresses in Soybean (Glycine max).

Strigolactones (SLs) are the major plant hormones that play important roles in regulating organ development and environmental stress tolerance in plants. Even though the SL-related genes have been identified and well characterized in some plants, the information of SL-related genes in soybean is not fully established yet, especially in response to salt and alkaline stresses. In this study, we identified nine SL biosynthetic genes that include two D27, two CCD7, two CCD8, and three MAX1, as well as seven SL signaling genes that comprised two D14, two MAX2, and three D53 in the soybean genome. We found that SL biosynthetic and signaling genes are evolutionary conserved among different species. Syntenic analysis of these genes revealed their location on nine chromosomes as well as the presence of 10 pairs of duplication genes. Moreover, plant hormone and stress-responsive elements were identified in the promoter regions of SL biosynthetic and signaling genes. By using reverse transcription quantitative real-time PCR, we confirmed that SL genes have different tissue expressions in roots, stems, and leaves. The expression profile of SL biosynthetic and signaling genes under salt and alkaline stresses further confirmed the regulatory roles of SL biosynthetic and signaling genes under stress. In conclusion, we identified and provided valuable information on the soybean SL biosynthetic and signaling genes, and established a foundation for further functional analysis of soybean SL-related genes in response to salt and alkaline stresses.

A study of zederone for the inhibition on ovarian cancer cell proliferation through mTOR/p70s6K signalling pathway.

PURPOSE: To investigate the inhibitory effect of Zederone (Zed) on the proliferation of human ovarian cancer cell line SK-OV-3 (SKOV3) and to explore the possible mechanism. METHODS: Cell Counting Kit-8 (CCK-8) assay was performed to detect the inhibitory effect of different concentrations of Zed on the proliferation of SKOV3 cells; the effect of Zed on the morphology of SKOV3 cells was observed; flow cytometry was performed to investigate the effect of Zed on the cycle phase distribution of SKOV3 cells; Real-time fluorescence quantitative polymerase chain reaction (qRT-PCR) and western blot were performed to detect the effects of Zed on the expression of mTOR, p70s6K, p-PI3K and p-Akt at mRNA and protein level in SKOV3 cells, respectively. RESULTS: Zed could effectively inhibit the proliferation of SKOV3 cells in vitro and change cell morphology. Flow cytometry indicated that Zed arrested SKOV3 cells at G1 phase. qRT-PCR revealed that Zed downregulated the mRNA levels of mTOR and p70s6K. However, western blot demonstrated that Zed could downregulate the protein expressions of mTOR, and phosphorylated p70 S6 kinase (p-p70s6K) in SKOV3 cells, but had no significant influences on p-PI3K and p-Akt proteins. CONCLUSION: In conclusion, Zed can significantly inhibit the proliferation of human ovarian cancer SKOV3 cells, and this regulation may be mediated by the inhibition of mTOR/p70s6K signal pathway.

Genome-wide identification and characterization of the soybean SOD family during alkaline stress.

BACKGROUND: Superoxide dismutase (SOD) proteins, as one kind of the antioxidant enzymes, play critical roles in plant response to various environment stresses. Even though its functions in the oxidative stress were very well characterized, the roles of SOD family genes in regulating alkaline stress response are not fully reported. METHODS: We identified the potential family members by using Hidden Markov model and soybean genome database. The neighbor-joining phylogenetic tree and exon-intron structures were generated by using software MEGA 5.0 and GSDS online server, respectively. Furthermore, the conserved motifs were analyzed by MEME online server. The syntenic analysis was conducted using Circos-0.69. Additionally, the expression levels of soybean SOD genes under alkaline stress were identified by qRT-PCR. RESULTS: In this study, we identified 13 potential SOD genes in soybean genome. Phylogenetic analysis suggested that SOD genes could be classified into three subfamilies, including MnSODs (GmMSD1-2), FeSODs (GmFSD1-5) and Cu/ZnSODs (GmCSD1-6). We further investigated the gene structure, chromosomal locations and gene-duplication, conserved domains and promoter cis-elements of the soybean SOD genes. We also explored the expression profiles of soybean SOD genes in different tissues and alkaline, salt and cold stresses, based on the transcriptome data. In addition, we detected their expression patterns in roots and leaves by qRT-PCR under alkaline stress, and found that different SOD subfamily genes may play different roles in response to alkaline stress. These results also confirmed the hypothesis that the great evolutionary divergence may contribute to the potential functional diversity in soybean SOD genes. Taken together, we established a foundation for further functional characterization of soybean SOD genes in response to alkaline stress in the future.

Brains of rhesus monkeys display Aß deposits and glial pathology while lacking Aß dimers and other Alzheimer's pathologies.

Cerebral amyloid beta (Aß) deposits are the main early pathology of Alzheimer's disease (AD). However, abundant Aß deposits also occur spontaneously in the brains of many healthy people who are free of AD with advancing aging. A crucial unanswered question in AD prevention is why AD does not develop in some elderly people, despite the presence of Aß deposits. The answer may lie in the composition of Aß oligomer isoforms in the Aß deposits of healthy brains, which are different from AD brains. However, which Aß oligomer triggers the transformation from aging to AD pathogenesis is still under debate. Some researchers insist that the Aß 12-mer causes AD pathology, while others suggest that the Aß dimer is the crucial molecule in AD pathology. Aged rhesus monkeys spontaneously develop Aß deposits in the brain with striking similarities to those of aged humans. Thus, rhesus monkeys are an ideal natural model to study the composition of Aß oligomer isoforms and their downstream effects on AD pathology. In this study, we found that Aß deposits in aged monkey brains included 3-mer, 5-mer, 9-mer, 10-mer, and 12-mer oligomers, but not 2-mer oligomers. The Aß deposits, which were devoid of Aß dimers, induced glial pathology (microgliosis, abnormal microglia morphology, and astrocytosis), but not the subsequent downstream pathologies of AD, including Tau pathology, neurodegeneration, and synapse loss. Our results indicate that the Aß dimer plays an important role in AD pathogenesis. Thus, targeting the Aß dimer is a promising strategy for preventing AD.

Fuzzy DEA-based classifier and its applications in healthcare management.

Nonlinear fuzzy classification models have better classification performance than linear fuzzy classifiers. In many nonlinear fuzzy classification problems, piecewise-linear fuzzy discriminant functions can approximate nonlinear fuzzy discriminant functions. In this paper, we first build fuzzy classifier based on data envelopment analysis (DEA) for incremental separable fuzzy training data, which can be widely applied in the healthcare management with fuzzy attributes, then we apply the proposed fuzzy DEA-based classifier in the diagnosis of Coronary with fuzzy symptoms and the classification of breast cancer dataset with fuzzy disturbance. Numerical experiments show the proposed fuzzy DEA-based classifier is accurate and robust.

Survey of Machine Learning Techniques for Prediction of the Isoform Specificity of Cytochrome P450 Substrates.

BACKGROUND: Determination or prediction of the Absorption, Distribution, Metabolism, and Excretion (ADME) properties of drug candidates and drug-induced toxicity plays crucial roles in drug discovery and development. Metabolism is one of the most complicated pharmacokinetic properties to be understood and predicted. However, experimental determination of the substrate binding, selectivity, sites and rates of metabolism is time- and recourse- consuming. In the phase I metabolism of foreign compounds (i.e., most of drugs), cytochrome P450 enzymes play a key role. To help develop drugs with proper ADME properties, computational models are highly desired to predict the ADME properties of drug candidates, particularly for drugs binding to cytochrome P450. OBJECTIVE: This narrative review aims to briefly summarize machine learning techniques used in the prediction of the cytochrome P450 isoform specificity of drug candidates. RESULTS: Both single-label and multi-label classification methods have demonstrated good performance on modelling and prediction of the isoform specificity of substrates based on their quantitative descriptors. CONCLUSION: This review provides a guide for researchers to develop machine learning-based methods to predict the cytochrome P450 isoform specificity of drug candidates.

Medium spiny neurons of the anterior dorsomedial striatum mediate reversal learning in a cell-type-dependent manner.

The striatum has been implicated in the regulation of cognitive flexibility. Abnormalities in the anterior dorsomedial striatum (aDMS) are revealed in many mental disorders in which cognitive inflexibility is commonly observed. However, it remains poorly understood whether the aDMS plays a special role in flexible cognitive control and what the regulation pattern is in different neuronal populations. Based on the reversal learning task in mice, we showed that optogenetic activation in dopamine receptor 1-expressing medium spiny neurons (D1R-MSNs) of the aDMS impaired flexibility; meanwhile, suppressing these neurons facilitated behavioral performance. Conversely, D2R-MSN activation accelerated reversal learning, but it induced no change through neuronal suppression. The acquisition and retention of discrimination learning were unaffected by the manipulation of any type of MSN. Through bi-direct optogenetic modulation in D1R-MSNs of the same subject in a serial reversal learning task, we further revealed the function of D1R-MSNs during different stages of reversal learning, where inhibiting and exciting the same group of neurons reduced perseverative errors and increased regressive errors. Following D1R- and D2R-MSN activation in the aDMS, neuronal activity of the mediodorsal thalamus decreased and increased, respectively, in parallel with behavioral impairment and facilitation, but not as a direct result of the activation of the striatal MSNs. We propose that D1R- and D2R-MSN sub-populations in the aDMS exert opposing functions in cognitive flexibility regulation, with more important and complex roles of D1R-MSNs involved. Mental disorders with cognitive flexibility problems may feature an underlying functional imbalance in the aDMS' two types of neurons.

Protein-protein interface hot spots prediction based on a hybrid feature selection strategy.

BACKGROUND: Hot spots are interface residues that contribute most binding affinity to protein-protein interaction. A compact and relevant feature subset is important for building machine learning methods to predict hot spots on protein-protein interfaces. Although different methods have been used to detect the relevant feature subset from a variety of features related to interface residues, it is still a challenge to detect the optimal feature subset for building the final model. RESULTS: In this study, three different feature selection methods were compared to propose a new hybrid feature selection strategy. This new strategy was proved to effectively reduce the feature space when we were building the prediction models for identifying hotspot residues. It was tested on eighty-two features, both conventional and newly proposed. According to the strategy, combining the feature subsets selected by decision tree and mRMR (maximum Relevance Minimum Redundancy) individually, we were able to build a model with 6 features by using a PSFS (Pseudo Sequential Forward Selection) process. Compared with other state-of-art methods for the independent test set, our model had shown better or comparable predictive performances (with F-measure 0.622 and recall 0.821). Analysis of the 6 features confirmed that our newly proposed feature CNSV_REL1 was important for our model. The analysis also showed that the complementarity between features should be considered as an important aspect when conducting the feature selection. CONCLUSION: In this study, most important of all, a new strategy for feature selection was proposed and proved to be effective in selecting the optimal feature subset for building prediction models, which can be used to predict hot spot residues on protein-protein interfaces. Moreover, two aspects, the generalization of the single feature and the complementarity between features, were proved to be of great importance and should be considered in feature selection methods. Finally, our newly proposed feature CNSV_REL1 had been proved an alternative and effective feature in predicting hot spots by our study. Our model is available for users through a webserver: http://zhulab.ahu.edu.cn/iPPHOT/ .

Functional inactivation of dorsal medial striatum alters behavioral flexibility and recognition process in mice.

Deficits in behavioral flexibility and recognition memory are commonly observed in mental illnesses and neurodegenerative diseases. Abnormality of the striatum has been implicated in an association with the pathology of these diseases. However, the exact roles of striatal heterogeneous structures in these cognitive functions are still unknown. In the present study, we investigated the effects of suppressing neuronal activity in the dorsomedial striatum (DMStr) and nucleus accumbens core (NAcC) on reversal learning and novelty recognition in mice. In addition, the locomotor activity, anxiety-like behavior and social interaction were analyzed. Neuronal inactivation was performed by expressing lentivirus-mediated tetanus toxin (TeNT) in the target regions. The results showed that reversal learning was facilitated by neuronal inactivation in the DMStr but not the NAcC, which was attributable to accelerated extinction of acquired strategy but not to impaired memory retention. Furthermore, mice with NAcC inactivation spent more time exploring a novel object than a familiar one, comparable to control mice. In contrast, mice with DMStr inactivation exhibited no preference to a novel environment during the novel object or place recognition test. The DMStr mice also exhibited decreased anxiety level. No phenotypic effect was observed in the locomotion or social interaction in mice with either DMStr or NAcC inactivation. Altogether, these findings suggest that the DMStr but not the ventral area of the striatum plays a crucial role in learning and memory by coordinating spatial exploration as well as mediating information updating.

N-methyl-D-aspartate receptor-mediated glutamate transmission in nucleus accumbens plays a more important role than that in dorsal striatum in cognitive flexibility.

Cognitive flexibility is a critical ability for adapting to an ever-changing environment in humans and animals. Deficits in cognitive flexibility are observed in most schizophrenia patients. Previous studies reported that the medial prefrontal cortex-to-ventral striatum and orbital frontal cortex-to-dorsal striatum circuits play important roles in extra- and intra-dimensional strategy switching, respectively. However, the precise function of striatal subregions in flexible behaviors is still unclear. N-methyl-D-aspartate receptors (NMDARs) are major glutamate receptors in the striatum that receive glutamatergic projections from the frontal cortex. The membrane insertion of Ca(2+)-permeable α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid receptors (AMPARs) depends on NMDAR activation and is required in learning and memory processes. In the present study, we measured set-shifting and reversal learning performance in operant chambers in rats and assessed the effects of blocking NMDARs and Ca(2+)-permeable AMPARs in striatal subregions on behavioral flexibility. The blockade of NMDARs in the nucleus accumbens (NAc) core by AP5 impaired set-shifting ability by causing a failure to modify prior learning. The suppression of NMDAR-mediated transmission in the NAc shell induced a deficit in set-shifting by disrupting the learning and maintenance of novel strategies. During reversal learning, infusions of AP5 into the NAc shell and core impaired the ability to learn and maintain new strategies. However, behavioral flexibility was not significantly affected by blocking NMDARs in the dorsal striatum. We also found that the blockade of Ca(2+)-permeable AMPARs by NASPM in any subregion of the striatum did not affect strategy switching. These findings suggest that NMDAR-mediated glutamate transmission in the NAc contributes more to cognitive execution compared with the dorsal striatum.