Construction of a chromosome-level genome assembly for genome-wide identification of growth-related quantitative trait loci in Sinocyclocheilus grahami (Cypriniformes, Cyprinidae).

The Dianchi golden-line barbel, Sinocyclocheilus grahami (Regan, 1904), is one of the "Four Famous Fishes" of Yunnan Province, China. Given its economic value, this species has been artificially bred successfully since 2007, with a nationally selected breed (" S. grahami, Bayou No. 1") certified in 2018. For the future utilization of this species, its growth rate, disease resistance, and wild adaptability need to be improved, which could be achieved with the help of molecular marker-assisted selection (MAS). In the current study, we constructed the first chromosome-level genome of S. grahami, assembled 48 pseudo-chromosomes, and obtained a genome assembly of 1.49 Gb. We also performed QTL-seq analysis of S. grahami using the highest and lowest bulks (i.e., largest and smallest size) in both a sibling and random population. We screened two quantitative trait loci (QTLs) (Chr3, 14.9-39.1 Mb and Chr17, 4.1-27.4 Mb) as the major growth-related locations. Several candidate genes (e.g., map2k5, stat1, phf21a, sox6, and smad6) were also identified, with functions related to growth, such as cell differentiation, neuronal development, skeletal muscle development, chondrogenesis, and immunity. These results built a solid foundation for in-depth MAS studies on the growth traits of S. grahami.

[Traditional Chinese Medicine data management policy in big data environment].

As traditional data management model cannot effectively manage the massive data in traditional Chinese medicine(TCM) due to the uncertainty of data object attributes as well as the diversity and abstraction of data representation, a management strategy for TCM data based on big data technology is proposed. Based on true characteristics of TCM data, this strategy could solve the problems of the uncertainty of data object attributes in TCM information and the non-uniformity of the data representation by using modeless properties of stored objects in big data technology. Hybrid indexing mode was also used to solve the conflicts brought by different storage modes in indexing process, with powerful capabilities in query processing of massive data through efficient parallel MapReduce process. The theoretical analysis provided the management framework and its key technology, while its performance was tested on Hadoop by using several common traditional Chinese medicines and prescriptions from practical TCM data source. Result showed that this strategy can effectively solve the storage problem of TCM information, with good performance in query efficiency, completeness and robustness.