China covers a vast geographic area, extending across the Palaearctic and Oriental regions. There are abundant species of Hymenoptera. For example bees, among 7 families and 21 subfamilies over the world, there are 69 genera, 14 subfamilies and 6 families in China. As we know, bees are reconized as most important pollinated insect groups. Prof. Yan-Ru WU ever estimated there might be over 3000 species in China, who recorded 576 species or subspecies of bees in China and published “Fauna of China-Mellitidae and Apidae” and “Fauna of China-Megachilidae” respectively in 2000 and 2006. In Chalcidoidea, most species of Eulophidae and Encyrtidae are potential natural enemies that can be used in biological control.

In China, there are lots of species yet to be discovered in any specific region. For the identified species, not all of them distribute in wide areas, and there are many special species in certain regions. Species identification has long been dependent on morphological characters. However, morphology is not good enough for closely related species. In the last decade, molecular sequence has been of much importance for taxonomical and phylogenetic studies. With sequenced data and those downloaded from public databases, we employed the DNA taxonomy for species identification, overcoming the disadvantages coming from morphological taxonomy. In practice, different gene regions may be sequenced by different labs. Thus, it could be the problematic for us to guarantee the consistence of different gene regions in order to construct the data matrix.

Till now, the General Mixed Yule Coalescent method (GMYC, Pons et al., 2006) has been widely used. Species identification based on GMYC hinges on pattern of branches, other than other information. Thus, it has large potential to identify species, and is especially useful for cryptic species. Though there have been available studies on automatic species identification or clustering, their limitations are obvious: 1) they were designed for single gene; 2) with the same criterion, results from different genes are contradictory to each other; 3) the CO1 gene was mainly used for animal DNA barcoding, however there is large amount of other gene sequences. So, a species clustering method is still needed, which could integrate multiple genes for constructing large matrix. This kind of clustering method has been programed by us, and is applying to more animal groups and larger matrix constructing. It is different from the method for integrating morphological characters which has less progresses recently. With the reference of Geniola et al., (2012), we would put forward it to cater further research.

Till now, molecular phylogenetics has enhanced our understanding of life evolution. The premise for molecular phylogenetics is the matrix of molecular characters, where the rows represent population, species, or other taxon level, and the columns represents homologous sites of sequences. Premier study suggests that adding rows has more influence on computation thatn adding columns. On the other hand, many studies manifest that adding columns results in better reliability of phylogenies. Nowadays, molecular phylogenetics tends to use multiple loci and even genomic data to reconstruct phylogenies, instead of one or two genes.

Facing the use of large data matrix, routine methods of molecular phylogenetics were challenged. Maximum likelihood (ML) and Bayesian inference (BI) have dominated in this field, since they employ mathematical models to infer evolutionary relationships. For ML, traditional software like PAUP has encountered bottleneck while computing large data; however, newly coming software such as RAxML and GARLI has solved this problem to some extent. What is more, RAxML could employ complex models and make corrections for heterogeneity among different genes, enhancing the reliability of algorithm. As for BI, available packages such as MrBayes and BayesPhylogenies could concern heterogeneity among different genes. But, it needs much computation time and efficient computers for large data. Out study suggests that efficiency of BI could still be improved via parallel computation. We would devote to paralleling the software of MrBayes to improve its efficiency for large data.

Orthologous genes form the basis for analyses of molecular phylogenetics. Recently, next-generation-sequencing techniques have emerged, enhancing the deep and efficiency of sequencing. They provide large data for molecular phylogenetics. However, assembling these small reads of large amount is a costly project. So, their application to molecular phylogenetics is relatively limited. Relatively, transcriptome is smaller and has less repetitive sequence units. The most expressed genes tend to be conservative because of their biological functions. It is probable that more orthologous genes would be sequenced in transcriptome data via next-generation sequencing. Transcriptome sequencing would be implemented on some groups to obtain large data for phylogenetic and evolutionary study.

Genomed-based phylogenetics would give reliable phylogeny and enhance molecular phylogenetics with large data matrix. At this moment, genome sequencing on each species is not enough. In the meantime, with more sequenced genomes, there are challenges for the available algorithms or software in molecular phylogenetics. While studying on high levels of phylogenetic relationships with available hardware and software conditions, we need investigations on genome sequencing of different species from the same genus, programing new and efficient algorithms, and updating and paralleling available algorithms or software.

It is the new ideas and methods that molecular taxonomy and systematics bring new insights into the systematics of Hymenoptera. However, the amounts of studies on molecular taxonomy of Hymenoptera remain fragmentary. There are not many molecular data of bees or chalcids available in China. Many species in the superfamily Chalcidoidea are parasitic, which can be used as natural enemies of insect pests. Most species of the superfamily Apoidea are pollinators. They play great roles in pollinating plants. But there are still some relatively important issues on the systematics about these insects: 1) It is difficult to realize systematic collection, taxonomies and identifications for multiple taxa owning to lack of talent scholars; 2) Probably only half of the bees are documented, while more undiscovered chalcids species remain yet to discover; 3) Classification on the level of subgenus or genus is still confounded for most species, while it is imperative to integrate many kinds of data for systematics; 4) With integrate new ideas and technologies to improve efficiency, current methods might meet the demands of agriculture and forestry; 5) Massive data including genomic data are accumulated and need to be integrated and analyzed; 6) Biology information of most species are unknown. All the above issues can be resolved by integrating taxonomy and molecular data on the basis of systematic theories and methods. So employing new molecular methods should be regarded as the main emphasis on systematic studies of China Hymenoptera, which will allow systematics make an important step forward.

Focusing on Hymenoptera species diversity in China and issues on morphology taxonomy, our team intends to integrate molecular taxonomy with multiple methods, to explore species diversity mechanisms, basing on the systematic collections and morphological identification.