Performance Comparison of Two Efficient Genomic Selection Methods(gsbay & MixP ) Applied in Aquacultural Organisms

Genomic selection is more and more popular in animal and plant breeding industries all around the world, as it can be applied early in life without impacting selection candidates. The objective of this study was to bring the advantages of genomic selection to scallop breeding. Two different genomic selection tools Mix P and gsbay were applied on genomic evaluation of simulated data and Zhikong scallop(Chlamys farreri) field data. The data were compared with genomic best linear unbiased prediction(GBLUP) method which has been applied widely. Our results showed that both Mix P and gsbay could accurately estimate single-nucleotide polymorphism(SNP) marker effects, and thereby could be applied for the analysis of genomic estimated breeding values(GEBV). In simulated data from different scenarios, the accuracy of GEBV acquired was ranged from 0.20 to 0.78 by Mix P; it was ranged from 0.21 to 0.67 by gsbay; and it was ranged from 0.21 to 0.61 by GBLUP. Estimations made by Mix P and gsbay were expected to be more reliable than those estimated by GBLUP. Predictions made by gsbay were more robust, while with Mix P the computation is much faster, especially in dealing with large-scale data. These results suggested that both algorithms implemented by Mix P and gsbay are feasible to carry out genomic selection in scallop breeding, and more genotype data will be necessary to produce genomic estimated breeding values with a higher accuracy for the industry.     

Performance comparison of two efficient genomic selection methods (gsbay & MixP) applied in aquacultural organisms

Genomic selection is more and more popular in animal and plant breeding industries all around the world, as it can be applied early in life without impacting selection candidates. The objective of this study was to bring the advantages of genomic selection to scallop breeding. Two different genomic selection tools MixP and gsbay were applied on genomic evaluation of simulated data and Zhikong scallop (Chlamys farreri) field data. The data were compared with genomic best linear unbiased prediction (GBLUP) method which has been applied widely. Our results showed that both MixP and gsbay could accurately estimate single-nucleotide polymorphism (SNP) marker effects, and thereby could be applied for the analysis of genomic estimated breeding values (GEBV). In simulated data from different scenarios, the accuracy of GEBV acquired was ranged from 0.20 to 0.78 by MixP; it was ranged from 0.21 to 0.67 by gsbay; and it was ranged from 0.21 to 0.61 by GBLUP. Estimations made by MixP and gsbay were expected to be more reliable than those estimated by GBLUP. Predictions made by gsbay were more robust, while with MixP the computation is much faster, especially in dealing with large-scale data. These results suggested that both algorithms implemented by MixP and gsbay are feasible to carry out genomic selection in scallop breeding, and more genotype data will be necessary to produce genomic estimated breeding values with a higher accuracy for the industry.     

���ģ��Ԥ�����������Ӳ�

??????????????????й????У????????????????????????????????????????????????????????仯???о??????????5???????????????????????ж?????????????????????????????????????????????????Щ??????????????????????????????????????????????????????3????????????????IGS????????????????????ζ???????????????????????????????????????????о???????3???????????Ч?????????????????????????????????????У??????????????????????????????編?????     

�����б�ѩ�����ʽģ�ͽ���ʱ��Ԥ�����о�

????????????????????????????????????б????????????????IGS???????????????????3???????????GPS??????????????1??????????????????????Ч?????з??????????????????????б?????????????????????????????????????????Ч??????С???????????б???????????????????????????????????????????     

抗菌肽对凡纳滨对虾生长和机体免疫的影响

基础饲料(对照组)中添加10g/kg抗菌肽制剂喂养体重0.10±0.01g、体长2.17±0.11cm的凡纳滨对虾(Penaeus vannamei)30d,进行10d的水浴攻毒(鳗弧菌10~8/mL),探讨饲料中添加抗菌肽对凡纳滨虾生长、成活率、免疫保护率,以及肌肉中溶菌酶活性、超氧化物歧化酶活性、总抗氧化能力的影响。结果表明:饲料中添加抗菌肽可显著提高凡纳滨对虾相对增重率、成活率和对致病菌的免疫保护率(P<0.05),但对饲料系数无显著影响(P>0.05);未受外界微生物刺激时,饲料中添加抗菌肽对凡纳滨对虾总抗氧化能力有促进作用(P<0.05),受外界微生物刺激后,饲料中添加抗菌肽喂养凡纳滨对虾可提高对虾超氧化物歧化酶活性(P<0.05)和总抗氧化能力(P<0.01)。因此饲料中添加抗菌肽喂养凡纳滨对虾可提高其机体非特异性免疫。     

优化循环神经网络在滑坡位移预测中的应用

以中国典型黄土滑坡域甘肃黑方台党川6#滑坡体为例,基于滑坡体北斗和位移计时序监测数据,首先利用深度学习框架Tensorflow分别构建3种循环神经网络滑坡位移预测模型：简单循环神经网络(simple recurrent neural network,SimpleRNN)、长短期记忆网络(long short-term memory,LSTM)和门控循环单元(gated recurrent unit,GRU),并进一步针对循环神经网络在参数设置时多采用经验手动调参或采用网格搜索法,易造成人为主观影响较大和计算效率低下的突出问题,引入遗传算法（genetic algorithm,GA）优化循环神经网络参数的自动最佳化选取,分别构建3种基于遗传算法改进的循环神经网络滑坡位移高精度预测模型：GA-SimpleRNN、GA-LSTM、GA-GRU。研究结果表明,改进参数自动寻优后的3种循环神经网络预测模型具有更优的预测性能,特别是GA-GRU模型预测精度最高,更适用于滑坡体长时序位移的高精度预测。     

一种区域精密天顶对流层延迟组合预报模型

针对现有区域天顶对流层延迟(ZTD)模型属于函数或格网型,参数固定,且难以表达ZTD时空快速变化特性等问题,提出一种基于小波变换、傅里叶级数拟合、自回归(AR)、支持向量回归(SVR)的组合预报新模型构建方法。该模型在时域内对ZTD序列进行小波变换,分解出低频和高频序列。低频序列采用傅里叶级数拟合成时间函数,高频序列则由AR进行预报。在空间域内利用SVR建立位置参数向傅里叶级数参数的映射。在该模型中输入时间与位置信息即可获取ZTD预报值。利用94个GNSS基站2 a的ZTD数据进行建模,24个GNSS基站1 a的ZTD数据进行预测对比。结果表明,实测值与模型预报值之间的平均偏差为-2.02 mm,均方根误差为3.07 cm,优于大部分区域ZTD模型。在伪距单点定位测试中,该模型能够显著提高定位精度。实验表明,该组合模型具有较高的预报精度和可靠性,具有一定的应用价值。     

人工神经网络和决策树模型在滑坡易发性分析中的性能对比

机器学习模型广泛应用于区域性滑坡易发性分析。模型的选择关系到评价结果的可信度、准确率和稳定性。现有滑坡易发性分析模型对比研究侧重模型的预测精度。模型的稳定性和数据量敏感性对机器学习模型的性能评估同样非常重要。本文以福建省南平市蔡源流域为研究区,以四川省绵阳市北川县为验证区,从预测精度、稳定性和数据量敏感性3个方面深入对比BP（Back Propagation）人工神经网络模型和CART（Classification and Regression Tree）决策树模型在滑坡易发性分析中的效果,主要结论如下:① 在逐渐增加一定数量训练样本的过程中,BP人工神经网络模型预测精度的增长率更高。在蔡源流域内,当训练样本数量增加10 000时,BP人工神经网络模型的预测精度上升5.22%,CART决策树模型的预测精度上升2.11%。② BP人工神经网络的预测精度高于CART决策树模型,且较为稳定。在100组数据集上,BP人工神经网络模型验证集预测精度的均值和验证集滑坡样本预测精度的均值分别为81.60%和84.86%,高于CART决策树模型的72.97%和76.59%。与此同时,BP人工神经网络模型对应预测精度的标准差分别是0.32%和0.37%,小于CART决策树模型的0.35%和0.67%。③ BP人工神经网络模型分析的滑坡易发区相比CART决策树模型,更接近实际滑坡的空间分布。最后,北川县的验证实验也出现了相同的现象。     

导航卫星钟差预报的Elman神经网络算法研究

针对卫星钟差不能被精确模型化的问题,将具有较强记忆功能和强大计算能力的Elman神经网络运用到卫星钟差预报中,提出适用于卫星钟差预报的Elman模型。首先对原始钟差数据进行一次差处理,然后选择合适的神经网络结构建立预报效果最佳的Elman钟差预报模型,最后选用国际GNSS服务（IGS）提供的精密钟差数据进行GPS卫星钟差预报,并与二次多项式模型、附加周期项的多项式模型和灰色系统模型进行对比分析。结果表明,Elman模型进行1 d、7 d和30 d钟差预报的精度得到显著提高,分别达到亚ns、ns和μs级,表明该模型的钟差预报性能优于3种常用模型,在卫星钟差预报中具有可行性。     

马氏珠母贝选系F_2早期选择反应和现实遗传力估计

对马氏珠母贝基础群体进行持续选择建立选系F2,比较了选系F2与对照群体早期生长差别,并估计了选择反应和现实遗传力。结果表明,在第8、14、21和35天,选系F2的平均壳长显著大于对照群体(P<0.05)。在第8、14、21和35天,选择反应和现实遗传力变化范围分别为0.63～0.89和0.36～0.51。本研究表明利用群体选择可以明显改良马氏珠母贝养殖群体的生长性状。     

马氏珠母贝壳长生长模型研究

以Cp统计量作为确定马氏珠母贝壳长最优生长模型的准则,通过Levenberg-Marquardt迭代法求出模型中各生长参数,用Brody、von Bertalanffy、Gompertz、Logistic和Richards 5个生长模型对马氏珠母贝壳长的生长进行了拟合。结果表明,在马氏珠母贝生长的第一年中,壳长生长过程遵循Brody生长模型,其壳长生长极限值为56.572 mm(43.807～69.337,95%置信区间)。     

马氏珠母贝经济性状对体重决定效应分析

采用通径分析方法研究了不同养殖区马氏珠母贝各分性状对体重的决定效应。结果表明:流沙养殖群体分性状壳长、壳高、壳宽、铰合线长、软体部重和闭壳肌重与体重间呈正相关(P<0.01),对体重的直接决定效应分别为0.091、0.072、0.029、0.002、0.089和0.001,间接决定效应分别为0.319、0.284、0.183、0.03、0.304和0.004,总决定效应分别为0.41、0.356、0.212、0.032、0.393和0.005;6个经济性状与体重间的关系在抽样断面上可用线性回归模型拟合,复相关系数达0.939(P<0.01)。6个性状3类决定效应按大小排列均为壳长>软体部重>壳宽>壳高>铰合线长>闭壳肌重。迈陈养殖群体作为验证群体,各效应值与前者有差异,但3类决定效应大小排序与前者完全相同。基于决定效应大小及性状易测性,育种目标性状的选择策略为首选壳长,二选壳高,三选壳宽。研究还发现壳宽对软体部重具有最大的决定效应,提出加大壳宽选择强度以培育大型珍珠新观点。     

高位池养殖凡纳滨对虾生长的研究

用高位池养殖凡纳滨对虾(Litopenaeus vannamei)进行生长研究。凡纳滨对虾养殖密度为135~165万尾/hm2,养殖海水盐度10.0~16.0,温度28.1~34.2℃,pH值7.6~8.8,DO5.0~6.0 mg/L。结果表明:凡纳滨对虾体长(L)与体重(W)的对应关系为:W=0.01 955L2.784;体长生长方程:Lt=13.82(1-е-0.012 01(t-0.699 4));体重生长方程:Wt=29.250 7×(1-е-0.012 01×(t-0.699 4))3;体长生长速率方程:dl/dt=13.823 1×0.012 01×е-0.01 20 1×(t-0.699 4);体重生长速率方程:dw/dt=3×29.250 7×0.012 01×е-0.012 01×(t-0.699 4)×(1-е-0.01201(t-0.699 4)2);高位池养殖凡纳滨对虾拐点年龄为91.36 d,拐点年龄前体长增长较快,拐点年龄后体长增长相对较慢,而拐点年龄前体重增长较慢,拐点年龄后体重增长较快。     

一种顾及钟差周期误差和随机特性的卫星钟差预报方法

提出一种顾及钟差周期误差和随机特性的卫星钟差预报方法。首先通过比较二次多项式加1、2、3、4个主要周期误差的模型,取其优者求得钟差预报的拟合值;然后针对拟合残差值的随机特性采用灰色模型进行建模,求得拟合值残差预报值;最后,将其与之前求得的预报值相结合得到最终的钟差预报值。采用IGS的15 min精密钟差数据进行实验,结果表明,在短期预报中,加2个主要周期误差的模型预报性能最好,并且新模型的预报精度优于常用算法。     

GROWTH PROMOTION OF RED SEA BREAM, PAGROSOMUS MAJOR, BY ORAL ADMINISTRATION OF RECOMBINANT EEL AND SALMON GROWTH HORMONE

ＩＮＴＲＯＤＵＣＴＩＯＮＧｒｏｗｔｈｈｏｒｍｏｎｅ (ＧＨ)ｉｓａｐｏｌｙｐｅｐｔｉｄｅｐｒｏｄｕｃｅｄｉｎｔｈｅｐｉｔｕｉｔａｒｙｇｌａｎｄａｎｄｈａｓｂｅｅｎｓｈｏｗｎｔｏｐｌａｙａｎｅｓｓｅｎｔｉａｌｒｏｌｅｉｎｔｈｅｓｔｉｍｕｌａｔｉｏｎｏｆｓｏｍａｔｉｃｇｒｏｗｔｈａｎｄｄｅｖｅｌｏｐｍｅｎｔｉｎｖｅｒｔｅｂｒａｔｅ .Ｓｅｖｅｒａｌｓｔｕｄｉｅｓｗｅｒｅｃｏｎｄｕｃｔｅｄｏｎｔｈｅｇｒｏｗｔｈ ｐｒｏｍｏｔｉｎｇａｃｔｉｖａｔｉｏｎｏｆｖａｒｉｏｕｓｔｅｌｅｏｓｔＧＨｓｉｎｓａｌｍｏｎｉｄ…     

基于GM(1,1)+AR模型的钟差短期预报改进算法研究

针对传统GM(1,1)+AR组合模型的缺点,提出一种可及时更新建模序列和增强数据间相关性的循环式钟差预报模型,在预报过程中根据预报时刻的不同实时调整AR模型阶数。考虑到原始钟差建模序列长度会对预报精度造成影响,分别使用2 h、6 h、12 h和24 h的钟差序列构建模型。实验结果表明,改进模型的预报精度较传统方法有一定提高,且预报结果更稳定;使用不同长度的钟差序列构建模型对预报结果有一定影响,其中二次多项式模型受原始序列长度的影响较大,改进模型受影响较小。     

基于LSTM神经网络的青藏高原月降水量预测

青藏高原的降水量预测不仅为该地区水资源合理规划利用提供依据,同时对中国及周边国家气候变化研究有着重要的意义。论文利用1990—2016年青藏高原降水量数据,采用长短期记忆神经网络（LSTM）对青藏高原月降水量进行预测,主要包括：① 使用青藏高原86个测站1990—2013年的月降水资料,预测各个测站2014—2016年的月降水量,并与传统的RNN、NAR、SSA和ARIMA预测模型相比,平均决定系数R²分别提高了0.07、0.15、0.13和0.36,均方根误差（RMSE）和平均绝对误差（MAE）表现更低;② 分析了降水量预测精度的空间分布特征,将各模型的R²在青藏高原地区内插值,分析R²的空间分布特征,发现所有模型降雨稀少的干旱地区和降雨多的湿润地区R²较低,在气候稳定、降水规律性明显的地区R²较高,且LSTM模型R²≥0.6的空间范围远大于传统模型;③ 分析了不同预测长度对各模型预测精度的影响,发现所有模型会随着预测长度增加而预测精度降低,但在不同的预测长度下LSTM预测的RMSE值都低于其他模型。     

ZResponse to selection, heritability and genetic correlations between body weight and body size in Pacific white shrimp, Litopenaeus vannamei

To quantify the response to selection,heritability and genetic correlations between weight and size of Litopenaeus vannamei,the body weight (BW),total length (TL),body length (BL),first abdominal segment depth (FASD),third abdominal segment depth (TASD),first abdominal segment width (FASW),and partial carapace length (PCL) of 5-month-old parents and of offspring were measured by calculating seven body measurings of offspring produced by a nested mating design.Seventeen half-sib families and 42 full-sib families ofL.vannamei were produced using artificial fertilization from 2-4 dams by each sire,and measured at around five months post-metamorphosis.The results show that heritabilities among various traits were high:0.515士0.030 for body weight and 0.394士0.030 for total length.After one generation of selection,the selection response was 10.70% for offspring growth.In the 5"’ month,the realized heritability for weight was 0.296 for the offspring generation.Genetic correlations between body weight and body size were highly variable.The results indicate that external morphological parameters can be applied during breeder selection for enhancing the growth without sacrificing animals for determining the body size and breed ability;and selective breeding can be improved significantly,simultaneously with increased production.     

恩诺沙星在凡纳滨对虾体内的代谢和残留消除规律

在26±2℃水温下,每天投喂含有恩诺沙星药物的饲料,研究恩诺沙星在凡纳滨对虾Litopenaeus vanname肌肉、肠和肝胰脏组织中的代谢和残留消除规律。残留药物用乙腈提取,液相色谱串联质谱仪检测。结果表明:恩诺沙星在凡纳滨对虾体内可代谢为环丙沙星,对虾体内同时有恩诺沙星和环丙沙星两种药物残留;环丙沙星在肌肉、肝胰脏和肠组织中的消除时间分别为6、8、10 d,而恩诺沙星在这三组织中的消除时间则为12、14、16 d。建议把肠作为该药残留监控的靶组织,凡纳滨对虾的休药期不少于16 d。