Assessment of molecular markers and marker-assisted selection for drought tolerance in barley(Hordeum vulgare L.)

This review updates the present status of the field of molecular markers and marker-assisted selection(MAS),using the example of drought tolerance in barley.The accuracy of selected quantitative trait loci(QTLs),candidate genes and suggested markers was assessed in the barley genome cv.Morex.Six common strategies are described for molecular marker development,candidate gene identification and verification,and their possible applications in MAS to improve the grain yield and yield components in barley under drought stress.These strategies are based on the following five principles:(1)Molecular markers are designated as genomic‘tags’,and their‘prediction’is strongly dependent on their distance from a candidate gene on genetic or physical maps;(2)plants react differently under favourable and stressful conditions or depending on their stage of development;(3)each candidate gene must be verified by confirming its expression in the relevant conditions,e.g.,drought;(4)the molecular marker identified must be validated for MAS for tolerance to drought stress and improved grain yield;and(5)the small number of molecular markers realized for MAS in breeding,from among the many studies targeting candidate genes,can be explained by the complex nature of drought stress,and multiple stress-responsive genes in each barley genotype that are expressed differentially depending on many other factors.

Transferring Translucent Endosperm Mutant Gene Wx-mq and Rice Stripe Disease Resistance Gene Stv-bi by Marker-Assisted Selection in Rice (Oryza sativa)

A high-yielding japonica rice variety, Wuyunjing 7, bred in Jiangsu Province, China as a female parent was crossed with a Japanese rice variety Kantou 194, which carries a rice stripe disease resistance gene Stv-b＇ and a translucent endosperm mutant gene Wx-mq. From F2 generations, a sequence characterized amplified region （SCAR） marker tightly linked with Stv-b＇ and a cleaved amplified polymorphic sequence （CAPS） marker for Wx-mq were used for marker-assisted selection. Finally, a new japonica rice line, Ning 9108, with excellent agronomic traits was obtained by multi-generational selection on stripe disease resistance and endosperm appearance. The utilization of the markers from genes related to rice quality and disease resistance was helpful not only for establishing a marker-assisted selection system of high-quality and disease resistance for rice but also for providing important intermediate materials and rapid selection method for good quality, disease resistance and high yield in rice breeding.

Pyramiding the disease resistant genes to southern rust and stalk rot in maize(Zea mays L.) with marker-assisted selection

Southern corn rust(SCR) caused by Puccinia polysora Underw and maize stalk rot caused by Pythium inflatum Matthews(MSR-2) are two destructive diseases of maize(Zea mays L.) in China.Our previous studies indicated that maize inbred line Qi319 is highly resistant to SCR but susceptible to MSR-2,while inbred line 1145 is highly resistant to MSR-2 but susceptible to SCR.The SCR resistant gene(RppQ) in Qi319 and MSR-2 resistant gene(Rpi1) in 1145 have been mapped on chromosome 10 and 4 respectively.In this research,through marker-assisted selection(MAS) with the molecular markers,bnlg1937 tightly linked to Rpi1 and phi041 tightly linked to RppQ,pyramid breeding of the two kinds of disease resistant genes were carried out from the year of 2003 to 2007.Two homozygotic inbred lines of F5 generation,DR94-1-1-1 and DR36-1-1-1 were identified.MAS result suggested DR94-1-1-1 and DR36-1-1-1 contained the two resistance genes RppQ and Rpi1.Field inoculation tests confirmed their high resistance to the two diseases.In addition,field investigation indicated that the two selected inbred lines,particularly DR94-1-1-1,had excellent agronomic traits such as plant height,ear height and yield-relating traits including ear length,ear diameter,ear weight,kernels per ear,kernels per row and kernel weight per ear.The two selected inbred lines DR94-1-1-1 and DR36-1-1-1 can either be directly developed into commercial variety or used as immediate donors of SCR and MSR resistance breeding programs in maize.

Breeding Rice Restorer Lines with High Resistance to Bacterial Blight by Using Molecular Marker-Assisted Selection

Two bacterial blight （BB） resistance genes, Xa21 and Xa4, from IRBB24 were introduced into hybrid rice restorer line Mianhui 725, which is highly susceptible to BB, by using hybridization and molecular marker-assisted selection technology. Four homologous restorer lines were obtained through testing the R target genes with molecular markers and analyzing parental genetic background. Inoculation of the four lines and their hybrids with the specific strains of Xanthomonas oryzae pv. oryzae, P1, P6 and seven representative strains of Chinese pathotype, C Ⅰ -CⅦ, showed that all of the four lines and their hybrids were highly resistant and presented broad resistance-spectrum to BB. The hybrids of G46A / R207-2 displayed good agronomic characters and high yield potential, and R207-2 was named Shuhui 207.

Marker-assisted Selection of ZmC_4Ppc in Rice Breeding and Yield Trait Performances of Advanced Lines

The full-length of intact Zea mays gene for phosphoenolpyruvate carboxylase gene （ZmC4Ppc） is 6 781 bp. The products of PCR for this gene were not clear with poor repeatability, resulting in that it was difficult for marker-assisted selection （MAS） both in rice and maize. For selecting the markers for MAS, sequences presented only in maize rather than in rice were identified by BLAST, and used for primer design using Primer Premier 5.0. A pair of specific primer termed MRpc （Forward： 5＇ AAGCAGGGAAGCGAGACG 3＇, Reverse： 5＇ GATTGCCGCCAGCAGTAG 3＇） was used for selection of transformed rice, and ZmC4Ppc could be highly and constitutively expressed at each tested developmental stages in the transformed rice selected by using MRpc. Thus, MRpc was used for MAS of progenies carrying ZmC4Ppc gene in rice and some restorer lines with ZmC4Ppc （e.g. FPM881） derived from ZmC4Ppc-transformed Kitaake backcrossed with a restorer line Shuhui 881 were obtained. The analyses on genetic background, PEPCase activity, net photosynthetic rate, general combining ability （GCA） and specific combining ability （SCA） of FPM881 showed that similarity of genetic background reached above 95%, the PEPCase and net photosynthetic rate were higher than those of the control, and some of the progenies carrying ZmC4Ppc gene had better GCA and SCA for grain yield per plant, number of panicles per plant, and 1000-grain weight than those of the control. This suggested that the introduction of maize ZmC4Ppc gene via MAS and its stable expression could increase grain yield of rice and would likely provide a pathway for rice varietal improvement.

Breeding of R8012,a Rice Restorer Line Resistant to Blast and Bacterial Blight Through Marker-Assisted Selection

Genetic improvement is one of the most effective strategies to prevent rice from blast and bacterial blight （BB） diseases, the two most prevalent diseases jeopardizing rice production. Rice hybrids with dural resistance to blast and BB are needed for sustainable production of food. An incomplete diallele design resulted in 25 crosses between five blast and five BB resistant germplasm accessions. Only one pair of parents, DH146 ×TM487, showed polymorphism for all the markers to identify one blast resistance gene Pi25 and three BB resistance genes, Xa21, xa13 and xa5, thus it was used in the marker-assisted selection （MAS）. F2 individuals of DH146× TM487 were genotyped using flanking markers of RM3330 and sequence tagged site （STS） marker SA7 for Pi25. The resistant F2 plants with Pi25 were used for pyramiding BB resistance genes Xa21, xa13 and xa5 identified by the markers pTA248, RM264 and RM153, respectively in subsequent generations. Finally, after selection for agronomic traits and restoration ability among 12 pyramided lines, we acquired an elite restorer line, R8012 including all four target genes （Pi25＋Xa21＋xa13＋xa5）. Hybrid Zhong 9NR8012 derived from the selected line showed stronger resistance to blast and BB, and higher grain yield than the commercial checks uniformally in experimental plots, 2007 state-wide yield trial and 2008 nation-wide yield trial. This study provides a paradigmatic example to show that MAS is a practically feasible tool in effectively pyramiding multiple resistance genes. The resultant restoring line and its hybrid would play an important role in securing rice production in China.

Improvement of Resistance to Rice Blast in Zhenshan 97 by Molecular Marker-aided Selection

Fungi blast is one of the most serious diseases of rice worldwide. Breeding resistant varieties have been proved to be the most effective and economical means to control the disease. This paper describes the molecular marker-assisted selection (MAS) procedure for a broad-spectrum blast resistant gene Pi1 integrated into an elite hybrid maintainer line, Zhenshan 97. A simple sequence repeat (SSR) based on molecular marker-aided selection system for Pi1 segment was established. Using a backcross population and a blast isolate F1829, Pi1 gene was mapped on the top of chromosome 11 between markers RZ536 and RM144, with a distance of 9.7 cM and 6.8 cM, respectively. Seventeen families derived from the recurrent parent Zhenshan 97 were obtained with homozygous Pi1 gene. The background of the 17 families was identified with inter simple sequence repeat (ISSR) amplification, the highest recovery of the Zhenshan 97 genetic background was 97.01% after the assay of 167 polymorphic bands.

Effects of the Fhb1 gene on Fusarium head blight resistance and agronomic traits of winter wheat

The gene Fhb1 has been used in many countries to improve wheat Fusarium head blight(FHB) resistance. To make better use of this gene in the Yellow-Huai River Valleys Winter Wheat Zone(YHWZ), the most important wheat-producing region of China, it is desirable to elucidate its effects on FHB resistance and agronomic traits in different genetic backgrounds. Based on a diagnostic marker for Fhb1, six BC2 populations were developed by crossing dwarf-male-sterile(DMS)-Zhoumai 16 to three Fhb1 donors(Ningmai 9, Ningmai 13, and Jianyang 84) and backcrossing to Zhoumai 16 and Zhoumai16’s derivative cultivars(Lunxuan 136 and Lunxuan 13) using marker-assisted backcross breeding. The progenies were assessed for FHB resistance and major agronomic traits.The Fhb1 alleles were identified using the gene-specific molecular marker. The plants with the Fhb1-resistant genotype(Fhb1-R) in these populations showed significantly fewer infected spikelets than those with the Fhb1-susceptible genotype(Fhb1-S). When Lunxuan 136 was used as the recurrent parent, Fhb1-R plants showed significantly fewer infected spikelets per spike than Fhb1-R plants produced using Lunxuan 13 as the recurrent parent, indicating that the genetic backgrounds of Fhb1 influence the expression of FHB resistance. Fhb1-R plants from the DMS-Zhoumai 16/Ningmai 9//Zhoumai 16/3/Lunxuan 136 population showed the highest FHB resistance among the six populations and a significantly higher level of FHB resistance than the moderately susceptible control Huaimai 20. No significant phenotypic differences between Fhb1-R and Fhb1-S plants were observed for the eight agronomic traits investigated. These results suggest that it is feasible to improve FHB resistance of winter wheat withoutreducing yield potential by introgressing Fhb1 resistance allele into FHB-susceptible cultivars in the YHWZ.

Identification of a Resistance Gene bls1 to Bacterial Leaf Streak in Wild Rice Oryza rufipogon Griff

Bacterial leaf streak （BLS） of rice, caused by Xanthomonas oryzae pv. oryzicola （Xoc） is a worldwide destructive disease. Development of resistant varieties is considered to be one of the most effective and eco-friendly ways to control the disease. However, only a few genes/QTLs having resistance to BLS have been identified in rice until now. In the present study, we have identified and primarily mapped a BLS-resistance gene, blsl, from a rice line DP3, derived from the wild rice species Oryza rufipogon Griff. A BC2F2 （9311/DP3//9311） population was constructed to map BLS-resistance gene in the rice line DP3. The segregation of the resistant and susceptible plants in BCzFz in 1：3 ratio （Z2=0.009, Z20 05,1=3.84, P〉0.05）, suggested that a recessive gene confers BLS resistance in DP3. In bulked segregant analysis （BSA）, two SSR markers RM8116 and RM584 were identified to be polymorphic in resistant and susceptible DNA bulks. For further mapping the resistance gene, six polymorphic markers around the target region were applied to analyze the genotypes of the BC2F2 individuals. As a result, the BLS-resistant gene, designated as blsl, was mapped in a 4.0-cM region flanked by RM587 and RM510 on chromosome 6.

Molecular Improvement of Grain Weight and Yield in Rice by Using GW6 Gene

Molecular design breeding is one of straightforward approaches to break yield barriers in rice. In this study, GW6 gene for grain length and width from Baodali was transferred into an indica recurrent parent 9311 and a japonica variety Zhonghua 11 （ZH11） using marker-assisted backcross （MAB）. One and three introgression lines were selected for phenotypic analysis from 9311 and ZH11 genetic backgrounds, respectively. SSL-1, an improved 9311 near isogenic line with GW6 performed 11%, 19% and 6.7%higher of grain length, 1000-grain weight and single plant yield, respectively, as compared with 9311. All the three improved ZH11-GW6 lines, R1, R2 and R3, had more than 30% increase in grain weight and about 7%higher in grain yield. Seed plumpness of R1, R2 and R3 was improved synchronously because the three ZH11-GW6 lines contained GIF1 （Grain Incomplete Filling 1）, a dominant grain filling gene. Thus, GW6 has high potential in increasing the yield of inbred lines through MAB, making it an important genetic resource in super hybrid rice breeding. This study provides insights in the utilization of GW6 for large grain and high yield rice breeding via molecular design breeding.

Molecular and Physical Mapping of Powdery Mildew Resistance Genes and QTLs in Wheat: A Review

Wheat powdery mildew （Pro） is a major disease of wheat worldwide. During the past years, numerous studies have been published on molecular mapping of Pm resistance gene（s） in wheat. We summarized the relevant findings of 89 major re- sistance gene mapping studies and 25 quantitative trait loci （QTL） mapping studies. Major Pm resistance genes and QTLs were found on all wheat chromosomes, but the Pm resistance genes/QTLs were not randomly distributed on each chromosome of wheat. The summarized data showed that the A or B genome has more major Pm resistance genes than the D genome and chromosomes 1A, 2A, 2B, 5B, 5D, 6B, 7A and 7B harbor more major Pm resistance genes than the other chromosomes. For adult plant resistance （APR） genes/QTLs, B genome of wheat harbors more APR genes than A and D genomes, and chromo- somes 2A, 4A, 5A, 1B, 2B, 3B, 5B, 6B, 7B, 2D, 5D and 7D harbor more Pm resistance QTLs than the other chromosomes, suggesting that A genome except 1A, 3A and 6A, B genome except 4B, D genome except 1D, 3D, 4D, and 6D play an impor- tant role in wheat combating against powdery mildew. Furthermore, Pm resistance genes are derived from wheat and its rela- tives, which suggested that the resistance sources are diverse and Pm resistance genes are diverse and useful in combating against the powdery mildew isolates. In this review, four APR genes, Pm38/Lr34/Yr18/Sr57, Pm46/Lr67/Yr46/Sr55, Pm？/Lr27/Yr30/ SY2 and Pm39/Lr46/Yr29, are not only resistant to powdery mildew but also effective for rust diseases in the field, indicating that such genes are stable and useful in wheat breeding programmes. The summarized data also provide chromosome locations or linked markers for Pm resistance genes/QTLs. Markers linked to these genes can also be utilized to pyramid diverse Pm resis- tance genes/QTLs more efficiently by marker-assisted selection.

Genetic dissection of the powdery mildew resistance in wheat breeding line LS5082 using BSR-Seq

Powdery mildew of wheat is a destructive disease seriously threatening yield and quality worldwide.Comprehensive dissection of new resistance-related loci/genes is necessary to control this disease.LS5082 is a Chinese wheat breeding line with resistance to powdery mildew.Genetic analysis,using the populations of LS5082 and three susceptible parents(Shannong 29,Shimai 22 and Huixianhong),indicated that a single dominant gene,tentatively designated PmLS5082,conferred seedling resistance to different Blumeria graminis f.sp.tritici(Bgt)isolates.Bulked segregant RNA-Seq was carried out to map PmLS5082 and to profile differentially expressed genes associated with PmLS5082.PmLS5082 was mapped to a 0.7 cM genetic interval on chromosome arm 2BL,which was aligned to a 0.7 Mb physical interval of 710.3–711.0 Mb.PmLS5082 differs from the known powdery mildew(Pm)resistance genes on chromosome arm 2BL based on their origin,chromosome positions and/or resistance spectrum,suggesting PmLS5082 is most likely a new Pm gene/allele.Through clusters of orthologous groups and kyoto encyclopedia of genes and genomes analyses,differentially expressed genes(DEGs)associated with PmLS5082 were profiled.Six DEGs in the PmLS5082 interval were confirmed to be associated with PmLS5082 via qPCR analysis,using an additional set of wheat samples and time-course analysis postinoculation with Bgt isolate E09.Ten closely linked markers,including two kompetitive allele-specific PCR markers,were confirmed to be suitable for marker-assisted selection of PmLS5082 in different genetic backgrounds,thus can be used to detect PmLS5082 and pyramid it with other genes in breeding programs.

Soybean cyst nematode-resistance: Gene identification and breeding strategies

Soybean cyst nematode(SCN,Heterodera glycines Ichinohe)is the most economically damaging disease of soybean worldwide,and breeding host plant resistance is the most feasible option for SCN management.In this review,we summarise the progress made so far in identifying nematode-resistance genes,the currently available sources of resistance,possible mechanisms of SCN resistance and strategies for soybean breeding.To date,only two sources of SCN resistance have been widely used,from the accessions PI 88788 and Peking,which has resulted in a shift in SCN resistance and created a narrow genetic base for SCN resistance.These resistant germplasms for SCN are classified into two types according to their copy number variation in a 31-kb genomic region:PI 88788-type resistance requires high copy numbers of a rhg1 resistance allele(rhg1-b)and Peking-type resistance requires both low copy numbers of a different rhg1 resistance allele(rhg1-a)and a resistant allele at another locus,Rhg4.Resistance related to rhg1 primarily involves impairment of vesicle trafficking through disruption of soluble NSF-attachment protein receptor(SNARE)complexes.By contrast,resistance via Rhg4 involves disturbance of folate homeostasis at SCN feeding sites due to alteration of the enzymatic activity of serine hydroxymethyltransferase(SHMT).Other potential mechanisms,including plant defences mediated by salicylic acid(SA)and jasmonic acid(JA)signalling modulation,have also been suggested for SCN resistance.Indeed,genome-wide association studies(GWAS)have identified other candidate SCN resistance genes,such as Gm SNAP11.Although gene functional analysis in a transient expression system could increase the efficiency of candidate gene identification,information on novel genes and mechanisms for SCN resistance remains limited.Any beneficial candidate genes identified might,when fully exploited,be valuable for improving the efficiency of marker-assisted breeding and dissecting the molecular mechanisms underlying SCN resistance.

Pyramiding of Pi46 and Pita to improve blast resistance and to evaluate the resistance effect of the two R genes

Utilization of R（resistance） genes to develop resistant cultivars is an effective strategy to combat against rice blast disease. In this study, R genes Pi46 and Pita in a resistant accession H4 were introgressed into an elite restorer line Hang-Hui-179（HH179） using the marker-assisted backcross breeding（MABB） procedure. As a result, three improved lines（e.g., R1791 carrying Pi46 alone, R1792 carrying Pita alone and R1793 carrying both Pi46 and Pita） were developed. The three improved lines had significant genetic similarities with the recurrent parent HH179. Thus, they and HH179 could be recognized as near isogenic lines（NILs）. The resistance spectrum of the three improved lines, which was tested at seedling stage, reached 91.1, 64.7 and 97.1%, respectively. This was markedly broader than that of HH179（23.5%）. Interestingly, R1793 showed resistance to panicle blast but neither R1791 nor R1792 exhibited resistance at two natural blast nurseries. The results implied that the stacking of Pi46 and Pita resulted in enhanced resistance, which was unachievable by either R gene alone. Further comparison indicated that the three improved lines were similar to HH179 in multiple agronomic traits; including plant height, tillers per plant, panicle length, spikelet fertility, and 1 000-grain weight. Thus, the three improved lines with different R genes can be used as new sources of resistance for developing variety. There is a complementary effect between the two R genes Pi46 and Pita.

Pyramiding disease resistance genes in elite winter wheat germplasm for Western Canada

We report on pyramiding different disease resistance genes against fungal pathogens in Canadian winter wheat germplasm based on available DNA markers and gene sequences.Genetic resistance represents a safe, economical and ecological method for protecting plants, growers and the health of consumers. Major diseases of wheat on the Canadian Prairies are common bunt, rusts(leaf, stem and stripe) and Fusarium head blight. Over the years markers for resistance genes against these diseases have been identified and used by the international wheat community. We describe markers that we have used to pyramid different resistance genes and indicate their presence in Canadian winter wheat cultivars issued from the winter wheat breeding program at the Agriculture and Agri-Food Canada,Lethbridge Research and Development Centre, the only winter wheat breeding program in Western Canada actively delivering new varieties for all regions of the Canadian Prairies.The sources of resistance and identities of PCR primers and amplification conditions are indicated to enable the transfer and pyramiding of different resistance(R) genes to breeding lines. We conclude by reviewing new tools for identifying R genes in wheat and indicate how mutagenesis and gene editing can help future efforts to extend the protection offered by known R genes.

利用双向导入系群体检测遗传背景对耐盐QTL定位的影响

以优质粳稻品种Lemont与高产籼稻品种特青为亲本培育的高代双向回交导入系为材料,在温室140 mmolL-1 NaCl胁迫条件下定位影响苗期叶片盐害级别(SST)、幼苗存活天数(SDS)、地上部K+浓度(SKC)和地上部Na+浓度(SNC)及人工气候室条件下影响地上部K+、Na+浓度的QTL。双向导入系的大部分遗传背景与各自的受体亲本相同,其中Lemont背景导入系中轮回亲本Lemont的基因组平均占83.8%,特青背景导入系中轮回亲本特青基因组平均占88.9%。各耐盐相关性状在两个背景群体中均出现超亲分离,多数性状的频率分布呈相互重叠状态,表明双亲作为供体相互导入各耐盐性状基因的效应大致相当。两个背景导入系群体中分别检测到影响上述耐盐相关性状的QTL各18个,同一性状在两个背景导入系中未能检测到任何相同表达的QTL,表明耐盐QTL表达具有很强的遗传背景效应,同时也说明这些耐盐QTL的效应可能较小。温室和人工气候室两种环境下仅在特青背景导入系中检测到1个影响SKC的相同QTL,表明耐盐QTL与环境的互作非常明显。虽然双亲均表现中等感盐,但QTL定位结果表明双亲中都存在一些提高耐盐相关性状的有利等位基因。研究认为,利用分子标记技术挖掘"隐蔽"于育成品种中的耐盐基因,进一步利用分子标记辅助选择技术对这些非等位耐盐基因进行聚合,完全有可能提高育成品种的耐盐水平。

从肿瘤基因表达数据挖掘分类规则的研究

基于肿瘤基因表达数据,利用信息科学的方法和技术建立肿瘤预测分类模型,对肿瘤基因表达模式研究和肿瘤的诊断识别具有重要意义。本研究提出一种从肿瘤基因表达数据中直接挖掘分类规则建立肿瘤预测分类器的方法。该方法首先抽取实验样本集,分别找出标记肿瘤和正常组织样本的分类特征,由此生成可预测样本类别的分类规则,对每个未知类别样本,按照置信度最高原则,选择一个分类规则作为预测结构。本研究的实验数据来自Broad Institute的前列腺癌基因表达数据,实验结果显示该方法的预测精度在90%以上,且同时获得了大量结构透明的分类预测规则,表明本研究的方法是可行的和有效的。

改进的克隆选择算法与SPEA相结合的进化算法

为了使进化过程中子代的繁殖能够像生物繁殖那样继承进化信息,通过挖掘抗体中优秀决定基并生成记忆集、增加高斯变异、用变异抗体群中亲和度高的抗体按概率替换记忆抗体群中低亲和度抗体等策略,提出了一种改进的克隆选择算法(ICSA)。将ICSA与SPEA相结合,形成了一种改进的克隆选择算法与强度Pareto进化算法相结合的新型的进化算法(ICSA-SPEA)。ICSA-SPEA通过克隆选择替代选择、交叉、重组等遗传操作。用一组多目标0/1背包问题测试算法性能的统计结果表明,改进的算法可以有效保持种群多样性,具有良好的收敛精度与准确度。

重用抗体优良片断的免疫进化算法

基于克隆选择原理与算法,通过分析具体现象阐述了改进克隆选择算法的思想来源,设计了挖掘抗体中优秀决定基因并生成记忆集、封装优秀决定基片段、用变异抗体群中亲和度高的抗体按概率替换记忆抗体群中低亲和度抗体的方法,获得了重用抗体优良片断的克隆选择算法.借鉴强度Pareto进化算法的进化框架,提出了重用抗体优良片断的免疫进化算法.该算法通过克隆选择替代选择、交叉、重组等遗传操作.在一组0/1背包问题上的测试结果表明,所提出的算法可以有效保持种群多样性,获得较高质量的Pareto非劣解集.

宏基因组中可移动序列的精确检测问题研究

基因组组装是宏基因组分析的主要挑战之一。通常假设所有测序序列均来源于同一个基因组,微生物中非常活跃的可移动元件给这个前提假设提出了重大质疑。文章将该质疑抽象为可移动元件与宿主染色体之间的二分类问题,准确的二分类性能将进一步促进宏基因组学方面的研究。基于宏基因组测序数据的数值化特征,详细考察特征选择算法Relief F、卡方检验和Fisher判别t检验,并结合分类模型逻辑回归、极限学习机、支持向量机和随机森林,验证最优可移动元件检测模型的性能。实验结果表明,Relief F特征选择算法和随机森林分类算法的融合模型,使用100个特征即可正确分类95%以上的宏基因组测序数据,优于使用全部的690个特征。