A brief review of machine learning methods for RNA methylation sites prediction.

Thanks to the tremendous advancement of deep sequencing and large-scale profiling, epitranscriptomics has become a rapidly growing field. As one of the most important parts of epitranscriptomics, ribonucleic acid (RNA) methylation has been focused on for years for its fundamental role in regulating the many aspects of RNA function. Thanks to the big data generated in sequencing, machine learning methods have been developed for efficiently identifying methylation sites. In this review, we comprehensively explore machine learning based approaches for predicting 10 types of methylation of RNA, which include m6A, m5C, m7G, 5hmC, m1A, m5U, m6Am, and so on. Firstly, we reviewed three main aspects of machine learning which are data, features and learning algorithms. Then, we summarized all the methods that have been used to predict the 10 types of methylation. Furthermore, the emergent methods which were designed to predict multiple types of methylation were also reviewed. Finally, we discussed the future perspectives for RNA methylation sites prediction.

m5CPred-SVM: a novel method for predicting m5C sites of RNA.

BACKGROUND: As one of the most common post-transcriptional modifications (PTCM) in RNA, 5-cytosine-methylation plays important roles in many biological functions such as RNA metabolism and cell fate decision. Through accurate identification of 5-methylcytosine (m5C) sites on RNA, researchers can better understand the exact role of 5-cytosine-methylation in these biological functions. In recent years, computational methods of predicting m5C sites have attracted lots of interests because of its efficiency and low-cost. However, both the accuracy and efficiency of these methods are not satisfactory yet and need further improvement. RESULTS: In this work, we have developed a new computational method, m5CPred-SVM, to identify m5C sites in three species, H. sapiens, M. musculus and A. thaliana. To build this model, we first collected benchmark datasets following three recently published methods. Then, six types of sequence-based features were generated based on RNA segments and the sequential forward feature selection strategy was used to obtain the optimal feature subset. After that, the performance of models based on different learning algorithms were compared, and the model based on the support vector machine provided the highest prediction accuracy. Finally, our proposed method, m5CPred-SVM was compared with several existing methods, and the result showed that m5CPred-SVM offered substantially higher prediction accuracy than previously published methods. It is expected that our method, m5CPred-SVM, can become a useful tool for accurate identification of m5C sites. CONCLUSION: In this study, by introducing position-specific propensity related features, we built a new model, m5CPred-SVM, to predict RNA m5C sites of three different species. The result shows that our model outperformed the existing state-of-art models. Our model is available for users through a web server at https://zhulab.ahu.edu.cn/m5CPred-SVM .

Genetic structure of Mount Huang honey bee (Apis cerana) populations: evidence from microsatellite polymorphism.

BACKGROUND: The Mount Huang eastern honey bees (Apis cerana) are an endemic population, which is well adapted to the local agricultural and ecological environment. In this study, the genetic structure of seven eastern honey bees (A. cerana) populations from Mount Huang in China were analyzed by SSR (simple sequence repeat) markers. RESULTS: The results revealed that 16 pairs of primers used amplified a total of 143 alleles. The number of alleles per locus ranged from 6 to 13, with a mean value of 8.94 alleles per locus. Observed and expected heterozygosities showed mean values of 0.446 and 0.831 respectively. UPGMA cluster analysis grouped seven eastern honey bees in three groups. CONCLUSION: The results obtained show a high genetic diversity in the honey bee populations studied in Mount Huang, and high differentiation among all the populations, suggesting that scarce exchange of honey bee species happened in Mount Huang. Our study demonstrated that the Mount Huang honey bee populations still have a natural genome worth being protected for conservation.

Interspecific competitive effects and annual differences in predation of Empoasca onukii (Hemiptera: Cicadellidae) by wandering spiders.

Empoasca onukii Matsuda is an important pest widely distributed in tea areas in China, which greatly affects tea production and quality. The long-term use of chemical control measures will cause environmental pollution. To better utilize wandering spiders that have strong competitive effects on the predation of E. onukii, we conducted a path analysis between the populations of E. onukii and wandering spiders in 3 tea plantations in 2020 and 2021. The wandering spider species that had the greatest direct and indirect effects on the population of E. onukii were analyzed by the magnitude of the path coefficient. Then, a gray system analysis was conducted to determine the closeness of the populations of different wandering spiders to E. onukii by the magnitude of gray correlation coefficient. Finally, the competition coefficients were calculated to determine the competitiveness of wandering spiders. In addition, considering the influence of the number of E. onukii on the interspecific competition of wandering spiders, the gray correlation coefficient and competition coefficient were combined to derive the competition intensity index, which was used to analyze the competitiveness of wandering spiders in a comprehensive manner. The highest competition coefficients in 2020 and 2021 were found for Ebrechtella tricuspidata Fabricius (Araneae: Thomisidae) (X2, 0.5329) and Clubiona reichlini Fabricius (Araneae: Clubionidae) (X4, 0.8475), respectively. The magnitude of the competition intensity index showed that the most competitive wandering spider in 2020 and 2021 was E. tricuspidata (X2, 0.5692) and C. reichlini (X4, 0.8892), respectively. The least competitive spider in both years was Plexippus setipes Karsch (Araneae: Salticidae) (X7). The more competitive the wandering spider is, the closer it is to E. onukii in terms of numbers, and the more dominant it is in feeding on E. onukii. By reasonably protecting and utilizing the competitive E. tricuspidata (X2) and C. reichlini (X4), we can achieve sustainable and effective control of E. onukii.

MTDeepM6A-2S: A two-stage multi-task deep learning method for predicting RNA N6-methyladenosine sites of Saccharomyces cerevisiae.

N6-methyladenosine (m6A) is one of the most important RNA modifications, which is involved in many biological activities. Computational methods have been developed to detect m6A sites due to their high efficiency and low costs. As one of the most widely utilized model organisms, many methods have been developed for predicting m6A sites of Saccharomyces cerevisiae. However, the generalization of these methods was hampered by the limited size of the benchmark datasets. On the other hand, over 60,000 low resolution m6A sites and more than 10,000 base resolution m6A sites of Saccharomyces cerevisiae are recorded in RMBase and m6A-Atlas, respectively. The base resolution m6A sites are often obtained from low resolution results by post calibration. In view of these, we proposed a two-stage deep learning method, named MTDeepM6A-2S, to predict RNA m6A sites of Saccharomyces cerevisiae based on RNA sequence information. In the first stage, a multi-task model with convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) deep framework was built to not only detect the low resolution m6A sites but also assign a reasonable probability for the predicted site. In the second stage, a transfer-learning strategy was used to build the model to predict the base resolution m6A sites from those low resolution m6A sites. The effectiveness of our model was validated on both training and independent test sets. The results show that our model outperforms other state-of-the-art models on the independent test set, which indicates that our model holds high potential to become a useful tool for epitranscriptomics analysis.

BERT-Kgly: A Bidirectional Encoder Representations From Transformers (BERT)-Based Model for Predicting Lysine Glycation Site for Homo sapiens.

As one of the most important posttranslational modifications (PTMs), protein lysine glycation changes the characteristics of the proteins and leads to the dysfunction of the proteins, which may cause diseases. Accurately detecting the glycation sites is of great benefit for understanding the biological function and potential mechanism of glycation in the treatment of diseases. However, experimental methods are expensive and time-consuming for lysine glycation site identification. Instead, computational methods, with their higher efficiency and lower cost, could be an important supplement to the experimental methods. In this study, we proposed a novel predictor, BERT-Kgly, for protein lysine glycation site prediction, which was developed by extracting embedding features of protein segments from pretrained Bidirectional Encoder Representations from Transformers (BERT) models. Three pretrained BERT models were explored to get the embeddings with optimal representability, and three downstream deep networks were employed to build our models. Our results showed that the model based on embeddings extracted from the BERT model pretrained on 556,603 protein sequences of UniProt outperforms other models. In addition, an independent test set was used to evaluate and compare our model with other existing methods, which indicated that our model was superior to other existing models.

A comprehensive comparison and analysis of computational predictors for RNA N6-methyladenosine sites of Saccharomyces cerevisiae.

N6-methyladenosine (m6A) modification, as one of the commonest post-transcription modifications in RNAs, has been reported to be highly related to many biological processes. Over the past decade, several tools for m6A sites prediction of Saccharomyces cerevisiae have been developed and are freely available online. However, the quality of predictions by these tools is difficult to quantify and compare. In this study, an independent dataset M6Atest6540 was compiled to systematically evaluate nine publicly available m6A prediction tools for S. cerevisiae. The experimental results indicate that RAM-ESVM achieved the best performance on M6Atest6540; however, most models performed substantially worse than their performances reported in the original papers. The benchmark dataset Met2614, which was used as the training dataset for the nine methods, were further analyzed by using a position bias index. The results demonstrated the significantly different bias of dataset Met2614 compared with the RNA segments around m6A sites recorded in RMBase. Moreover, newMet2614 was collected by randomly selecting RNA segments from non-redundant data recorded in RMBase, and three different kinds of features were extracted. The performances of the models built on Met2614 and newMet2614 with the features were compared, which shows the better generalization of models built on newMet2614. Our results also indicate the position-specific propensity-based features outperform other features, although they are also easily over-fitted on a biased dataset.

TBC2target: A Resource of Predicted Target Genes of Tea Bioactive Compounds.

Tea is one of the most popular non-alcoholic beverages consumed worldwide. Numerous bioactive constituents of tea were confirmed to possess healthy benefits via the mechanisms of regulating gene expressions or protein activities. However, a complete interacting profile between tea bioactive compounds (TBCs) and their target genes is lacking, which put an obstacle in the study of healthy function of tea. To fill this gap, we developed a database of target genes of TBCs (TBC2target, http://camellia.ahau.edu.cn/TBC2target) based on a pharmacophore mapping approach. In TBC2target, 6,226 interactions between 240 TBCs and 673 target genes were documented. TBC2target contains detailed information about each interacting entry, such as TBC, CAS number, PubChem CID, source of compound (e.g., green, black), compound type, target gene(s) of TBC, gene symbol, gene ID, ENSEMBL ID, PDB ID, TBC bioactivity and the reference. Using the TBC-target associations, we constructed a bipartite network and provided users the global network and local sub-network visualization and topological analyses. The entire database is free for online browsing, searching and downloading. In addition, TBC2target provides a BLAST search function to facilitate use of the database. The particular strengths of TBC2target are the inclusion of the comprehensive TBC-target interactions, and the capacity to visualize and analyze the interacting networks, which may help uncovering the beneficial effects of tea on human health as a central resource in tea health community.

WaspBase: a genomic resource for the interactions among parasitic wasps, insect hosts and plants.

Insect pests reduce yield and cause economic losses, which are major problems in agriculture. Parasitic wasps are the natural enemies of many agricultural pests and thus have been widely used as biological control agents. Plants, phytophagous insects and parasitic wasps form a tritrophic food chain. Understanding the interactions in this tritrophic system should be helpful for developing parasitic wasps for pest control and deciphering the mechanisms of parasitism. However, the genomic resources for this tritrophic system are not well organized. Here, we describe the WaspBase, a new database that contains 573 transcriptomes of 35 parasitic wasps and the genomes of 12 parasitic wasps, 5 insect hosts and 8 plants. In addition, we identified long non-coding RNA, untranslated regions and 25 widely studied gene families from the genome and transcriptome data of these species. WaspBase provides conventional web services such as Basic Local Alignment Search Tool, search and download, together with several widely used tools such as profile hidden Markov model, Multiple Alignment using Fast Fourier Transform, automated alignment trimming and JBrowse. We also present a collection of active researchers in the field of parasitic wasps, which should be useful for constructing scientific networks in this field.

Stress2TF: a manually curated database of TF regulation in plant response to stress.

Considerable studies demonstrate that plant transcription factors (TFs) play key regulatory roles in abiotic/biotic stress conditions, such as drought and pathogen attack. However, there is no effort dedicated to curate experimentally validated stress-TF regulatory relationships from these individual reports into a central database, which put an obstacle in the exploration of stress-TF regulations in plants. To address this issue, we presented a literature-curated database 'Stress2TF' that currently documented 1533 regulatory relationships between 71 abiotic/biotic stresses and 558 TFs in 47 plant species. Each entry in Stress2TF contains detailed information about a stress-TF relationship such as plant name, stress name, TF and brief description of stress-TF relationship. Stress2TF provided a user-friendly interface for entry browse, search and download. In addition, a submission page and several useful tools (e.g., BLAST, network visualization) were integrated. Stress2TF may be a valuable resource for the research of stress-TF regulatory mechanisms in plants. Stress2TF is available at http://csgenomics.ahau.edu.cn/Stress2TF.

[Relationships of bee population fluctuation and distribution with natural environment in Anhui province].

In 2002 to approximately 2004, an investigation was made on the bee population dynamics and its relationships with the ecological environment in four ecological regions of Anhui Province. The results indicated that in the mountainous areas of south and west Anhui, there were 46 and 37 species of nectariferous plants, and the distribution density of Apis cerena cerena population was 2.01 and 1.95 colony x km(-2), respectively. In Jianghuai area and Huaibei plain, there were 17 and 12 species of nectariferous plants, which had concentrated and short flowering period and fitted for Apis mellifera Ligustica oysterring and producing, and the distribution density of Apis cerena cerena population was 0. 06 and 0. 02 colony x km(-2), respectively. Bee population fluctuation and distribution was affected by wasp predation. The breeding proportion of Apis cerena cerena to local apis population was 41.5%, 36.8%, 3.1% and 1.1%, and that of Apis mellifera Ligustica was 58.5%, 63.2%, 96.9% and 98.9% in the mountainous areas of south and west Anhui, Jianghuai area, and Huaibei plain, respectively.

[Structural characteristics of arthropod community in grapery].

By the method of principal component analysis, this paper analyzed the characteristic parameters of arthropod community and its phytophagous and predacious sub-communities in grapery. The results showed that in the first principal component, the contribution of arthropod community, phytophagous sub-community and predacious sub-community was 66.70%, 73.39% and 54.17%, respectively. For arthropod community, the absolute values of normalized regressive coefficients of individuals' number N, Hill diversity index N1, Hill diversity index N2 and McIntosh index D(mc) were larger, indicating their greater contribution on the community. For phytophagous sub-community, N, N1, N2, N and S (total species) had greater contribution; while for predacious sub-community, the greater contribution was made by N1, N2, Shannon-Wiener diversity index H' and abundance R. Synthetically, N2 and N1 had greater contribution on the arthropod community and its phytophagous and predacious sub-communities in grapery, comparing with other characteristic parameters.

[Characteristics, dynamics and niche of insect community in plum orchard].

The insect community in plum orchard was investigated on organization level and temporal-spatial niche. The results showed that the insect community was abundant, which included 6 orders, 23 families. The individuals of species, diversity indices, and evenness increased with time. Myzus persicae and Asiaarposina sasokii had the widest spatial niche breadth, while Didesmococcus koreauus borchs had the widest temporal niche breadth. Among the natural enemies, Chilocorus rubidus had the widest both temporal and spatial niche breadth. The niche of Chicocorus rubidus and Didesmococcus koreauus overlapped larger than that of the others, which indicated their synchrony in temporal dimension and their similarity in spatial dimension. As the dominant natural enemies, the two populations should be protected and utilized to control plum pest.

[Geostatistics analysis on spatial patterns of Myzus persicae and Erigonidium graminicola in plum orchard].

Investigations on the spatial construction and distribution of Myzus persicae and Erigonidium graminicola in a plum orchard were conducted from March 2003 to November 2003. The results indicated that the semivariogram of Myzus persicae could be described by spherical model, except on June 27 and November 22, which should be described by lined model, and that of Erigonidium graminicola could be described by spherical model, except on May 21, May 31, October 19 and November 22, which should be described by lined model. It could be concluded that the amount and spatial distribution of Erigonidium graminicola was closely related to those of Myzus persicae.

[Spatial distribution pattern and space dependence of Contarinia sp. and Erigonidium graminicola in jujube orchard].

With the principles and methodologies of geostatistics, this paper studied the spatial structure and space dependence of Contarinia sp. and its natural enemy Erigonidium graminicola in a jujube orchard at different date. The results indicated that the semivariogram of Contarinia sp. on 8 June, 27 August and 19 October could be described by spherical model, showing an aggregated spatial distribution with the ranges of spatial dependence being 26.3, 7.8 and 22.0 m, and the degrees of spatial dependence being 88.05%, 85.77% and 87.58%, respectively. The semivariogram of Erigonidium graminicola could be also described by spherical model, showing an aggregated distribution with the ranges of spatial dependence being 28.3, 22.5 and 22.2 m, and the degrees of spatial dependence being 90.09%, 87.96% and 85.72%, respectively. It is suggested that Erigonidium graminicola is the dominant natural enemy against Contarinia sp. population in the examined jujube orchard.

[Effects of management level on community characteristics of arthropod and on population numbers of target insect pest and its natural enemies in graperies].

In this paper, an investigation on the grape tree and ground vegetation was conducted in two graperies with intensive and extensive management, aimed to study the effects of different management level on the community characteristics of arthropod, and the population numbers of target pest Halticinae chalybca (Illiger) and its natural enemies Erigonidium gram inicolum and Tetragnathidae. The results showed that between the two graperies, the individual number, concentration value, evenness, and Hill diversity index of arthropod community had no significant difference, but its species number and abundance was significantly different (P < 0.05). The species number of arthropod on the grape trees in intensive management grapery was not significantly different from that in extensive management grapery, while on the ground vegetation, it was significantly different (P < 0.05). There was a little difference in the population numbers of H. chalybca and its natural enemies on the trees and ground vegetations of the two graperies.

[Arthropod community structure and its fuzzy clustering analysis in jujube orchard].

Based on the principles of community ecology and by the method of fuzzy clustering, this paper studied the temporal and quantitative dynamics of the arthropod community and sub-community in jujube orchard (Feidong 2003), aimed to utilize natural enemy resources for pest control. There were 52 species of arthropod belonging to 47 families in the orchard. Arachnida and Cecidomyidae had a higher relative abundance than other families being 0.3465 and 0.2309, respectively. The individuals of the arthropod community were the highest on 8 June and 21 September. The total community and pests-neutral insect sub-community could be clustered into 4 clusters, respectively. For total community, the clusters (T = 0.92) were those on 22 March, 6 April, 22 November, and on other dates, and for pests-neutral insect sub-community, they (T = 0.94) were those on 22 March, 6 April, 8 June, and on other dates. The natural enemy sub-community could be clustered into 3 clusters (T = 0.94), i. e., those on 22 March, on 6 April, and on other dates. The results of cluster analysis partly reflected the seasonal differences of total community and sub-community, while the temporal overlaps of cluster results reflected the complexity of community structure.

[Cluster analysis on the temporal dynamics of arthropod community in a plum orchard].

In this paper, twelve investigations were conducted on the structural feature of arthropod community in a plum orchard, and a cluster analysis was made on the temporal dynamics of the community. The total community could be clustered into 5 clusters (D = 0.2000), i.e., that in March, in June, in July, in November, and in other months. The natural enemy and pest-neutral sub-communities could be also clustered into 5 clusters, respectively. For natural enemy sub-community, the clusters (D = 0.2000) were that in March, in July, in August, in September and October, and in other months, and for pest-neutral sub-community, they (D = 0.1000) were that in April, in July, in June, in November, and in other months. The results of cluster analysis partly reflected the seasonal differences of total community and sub-communities, while the temporal overlaps of cluster results reflected the complexity of community structure. Based on the optimization cut-apart, both the total community and the sub-communities were divided into 5 stages, i.e., 6 April as the first stage, 27 April to 8 June as the second stage, 27 June to 27 August as the third stage, 21 September to 19 October as the forth stage, and 22 November as the fifth stage.

[Relationship between occurrence of Dryocosmus kuriphilus and development of cecidum].

Dryocosmus kuriphilus is one of the most important pests on chestnut in China, and cecidum is the jeopardized result of Dryocosmus kuriphilus. Under natural condition, the growth status of different cecidum of Dryocosmus kuriphilus was differed, and the weight of individual cecidum was submitted to normal distribution. The zooecium number was positively related to cecidum quality and volume. For three different varieties of chestnut, their resistance to Dryocosmus kuriphilus were different, and the order was Chushuhong > Mifengqiu > Ershuizao. Torymus sinensis Kamijo was the dominant natural enemy. The cecidum-parasitized ratio was 72.7%, and the zooecium-parasitized ratio was 24.06%.

[Geostatistical analysis on spatial patterns of Didesmococcus koreanus Borchs and Chilocorus rubidus Hope].

The spatial pattern and correlation of Didesmococcus koreanus Borchs and Chilocorus rubidus Hope in different periods were investigated with geostatistics. The results showed that the semivariogram of Didesmococcus koreanus Borchs and Chilocorus rubidus Hope could be described by spherical model, indicating that their spatial pattern were in agreement. The spatial variation was arosed by its spatial autocorrelation in small scales, and the spatial pattern had no apparent changes in different periods. The range varied from 9.56-15.70 m and from 18.83-19.74 m, respectively. The amount and spatial distribution of Chilocorus rubidus Hope was closely related to that of Didesmococcus koreanus Borchs, which revealed that Chilocorus rubidus Hope was the dominant natural enemy of Didesmococcus koreanus Borchs.