connectivity网易轻博客
网易轻博客 时间:2021-01-13 阅读:()
1ElectronicSupplementaryMaterialGeneticAssessmentofEnvironmentalFeaturesthatInfluenceDeerDispersal:ImplicationsforPrion-InfectedPopulationsAmyC.
Kelly,NohraE.
Mateus-Pinilla,WilliamBrown,MarilynO.
Ruiz,MarlisR.
Douglas,MichaelE.
Douglas,PaulShelton,TomBeissel,JanNovakofskiMicrosatelliteMarkersThefollowingmicrosatelliteswereemployedinthisstudy:BM1225,BM4107,CSN3,(Bishopetal.
1994),IGF-1(Kirkpatrick1992),OBCAM(Friesetal.
1993),OarFcb304(Buchananetal.
1993),RT20,RT23,RT27(Wilsonetal.
1997)andSrcrsp-10(Bhebheetal.
1994).
Welabeledforwardprimerswithfluorescentdyes(NED,HEX,FAM)andseparatedmicrosatellitefragmentsonanABI3730XLcapillarysequencer(AppliedBiosystems,FosterCity,CA).
WevisualizedmicrosatellitegenotypeswithGeneMapper(v.
4.
0;AppliedBiosystems,FosterCity,CA).
WeusedMicro-checker(v.
2.
2.
3;VanOosterhoutetal.
2004)toevaluategenotypingerrorsusingexpectedallelefrequenciesderivedunderHardy-Weinbergequilibrium(HWE).
FSTSurfaceProjectionWeusedtheSingleSpeciesGeneticDivergenceoptionwithintheGeneticLandscapesGIS(GeographicInformationSystem)ToolboxtoprojectasurfacefrompairwiseFSTvaluescalculatedbetweenall31studysites.
TheprogramfirstassociatedpairwiseFST2valueswithmidpointsbetweenallstudysitesandanetworkofnearestneighbors.
Spatialinterpolationwasthenperformedusinganinversedistanceweightedinterpolationalgorithmtoestimategeneticdistancesalongagridoverlaidonthestudyarea.
GeneticdistancesforallpointsacrossthegridwereinterpolatedsuchthatmidpointFSTvaluesthatwerespatiallycloserinfluencedtheestimatemoresothanthosethatweredistant.
Moredetailsontheinterpolationprocedurearedescribedinhttp://www.
werc.
usgs.
gov/productdetails.
aspxid=4017.
FRAGSTATSmetricsTheConnectanceIndex(CONNECT)measuresfunctionalconnectivity,meaningthatgridcellsinthedatathatdepictthetargetvariablearenotliterallyadjacent,buttheyareconsideredadjacent(orconnected)withinagiventhresholddistance.
Inthiscase,adjacencywasdefinedascellswithin100mofeachother.
Theuser-defined100mthresholdwasusedtoaccountforpotentialimprecisionofdataclassificationsatfinespatialresolutionsandtoprovideamorerealistic(i.
e.
,functional)depictionofhowdeermightinteractwiththelandscape.
Themetricitselfisapercentage,witharangeof0to100.
Morespecifically,itmeasuresthepercentageoftargetvariableadjacencies(connectionsorjoins)relativetoallpossibleadjacencies.
FormoreinformationontheConnectanceIndexsee:http://www.
umass.
edu/landeco/research/fragstats/documents/Metrics/Connectivity%20Metrics/Metrics/C122%20-%20CONNECT.
htm3ThePatchCohesionIndex(COHESION)isasecondmeasureofconnectivityofalandscapevariable.
Thismetrictakesintoaccountphysicaladjacency(withoutathreshold)incombinationwiththesizeandshapeofthepatches.
Takingforestasanexample,ahigherCOHESIONvaluewouldoccurinalandscapewithlargerandcompactpatchescomparedtoonewithsmallorconvolutedpatches.
FormoreinformationonthePatchCohesionIndexsee:http://www.
umass.
edu/landeco/research/fragstats/documents/Metrics/Connectivity%20Metrics/Metrics/C121%20-%20COHESION.
htmTheClumpinessIndex(CLUMPY)isametricindicatinghowcontiguousordispersedaretheadjacentpatchesofalandscapevariable.
AhighervalueofCLUMPYwouldoccurifseveralpatcheswerelocatedclosetogetherratherthanbeingmoreuniformlydistributed.
FormoreinformationontheClumpinessIndex(CLUMPY)seehttp://www.
umass.
edu/landeco/research/fragstats/documents/Metrics/Contagion%20-%20Interspersion%20Metrics/Metrics/C115%20-%20CLUMPY.
htmThePerimeter-AreaFractalDimension(PAFRAC)isashapemetricdeterminedacrossarangeofspatialscales.
PARFRACislowforpatcheswithsimpleperimetersandincreasesforpatchshapeswithhighlyconvolutedperimeters.
FormoreinformationonthePerimeter-AreaFractalDimensionIndex(PAFRAC),seehttp://www.
umass.
edu/landeco/research/fragstats/documents/Metrics/Shape%20Metrics/Metrics/C23%20-%20PAFRAC.
htm.
Multivariatelinearregressionanalysis4DescriptionandsourceoflandscapevariablesincludedinmultivariateregressionanalysisarelistedinTableS1.
Topreventoverlyinfluentialobservationsfrombiasingourmodels,weusedleveragescores,Cook'sDvalues,andstandardizedinfluencevaluestoidentifyoutliers(Kieetal.
2002;ChatterjeeandHadi2009;Anlaufetal.
2011).
Leveragescoresidentifyobservationsthatresultinlargechangesinregressionlinefitupontheirdeletion.
Wecalculatedleverage(pi)accordingtoChatterjeeandHadi(1986)andconsideredobservationsoverlyinfluentialwhenpi>2p/N(p=numberofindependentvariablesinthemodel;N=numberofobservations).
Cook'sDvalueswerecalculatedaccordingtoCook(1977)andcomparedtoanFdistributionwithα=0.
05and(N-p)degreesoffreedom.
AllCook'sDvalues>thecriticalFvaluewereconsideredoverlyinfluentialandremovedfromthemodel(Cook1977).
LeveragescoresandCook'sDallowedustodeterminetheeffectsofoutliersontheoverallmodel,butstandardizedinfluencevalues(DFFITS)allowedustoexaminetheinfluenceofeachobservationonitspredictedvalue.
WecalculatedDFFITSaccordingtoChatterjeeandHadi(1986)andeliminatedobservationsyieldingvalues>2)/(Np(ChatterjeeandHadi1986).
Usingthesethreecriteria,weidentifiedthirteenobservationsoutof465(2.
8%)thatwereoutliersandafterstringentlyevaluatingtheirbasis(Motulsky2010),weomittedthemduringfurtheranalyses.
Themajorityoftheoutliersremoved(7/13)involvedstudysitesthathadrelativelylowsamplesizes.
Threeofthirteenoutliersinvolvedpairwisecomparisonswithstudysite27,thoughtheremainingtenoutliersappearedtoinvolvestudysitesthatwererandomlydistributedgeographically.
AsingleoutlierhadthehighestFSTvalueobserved,thoughtheremainingoutliersdidnotexhibitunusuallyhighorlowFSTvaluesascomparedtotherestofthe5dataset.
WecomparedvaluesofdependentvariablesofoutlierstovaluesfortherestofthedatabyexaminingboxplotsandplottingdependentvariablesagainstFSTvalues(datanotshown).
Trendsinthedistributionofvaluesfordependentvariablewerenotapparentinoutliersascomparedtotherestofthedata.
Whentwoormorelandscapevariableswerehighlycorrelated(Pearson'srP>0.
7),thepredictorwiththelowestpartialcorrelationinthefullmodelwasremoved.
RemovinglandscapevariableswithrP>0.
7(n=7)resultedinagenerallackofcollinearityamongpredictorsasdeterminedbyvarianceinflationfactors.
CorrelatedpredictorsthatwereremovedfromthemodelarelistedinTableS2.
Weusedvarianceinflationfactors(VIF)toevaluatetheincreaseinvarianceforestimatedregressioncoefficientsresultingfromcollinearpredictors,withVIF>10indicativeofhighmulticollinearity(Kutneretal.
2004).
Afterremovinghighlycorrelatedvariables,wecalculatedvarianceinflationfactorsforindependentvariablesandfoundthatthevarianceofestimatedregressioncoefficientswasnotsubstantiallyincreasedbycollinearpredictorsasVIFvaluesforallpredictorswere0.
7thatweresubsequentlyremovedfromthemodel.
VariableCorrelateDirectionofCorrelationVariableRemoved*%GrasslandSlope+%GrasslandForestCONNECTDevelopedCONNECT+ForestCONNECT%GrasslandGrasslandCONNECT-%GrasslandForestCONNECTGrasslandCONNECT+ForestCONNECTForestCONNECTWaterCONNECT+ForestCONNECTAgricultureCLUMPY%Agriculture-AgricultureCLUMPY%RiparianSlope+Slope%GrasslandForestCLUMPY-%GrasslandForestCONNECTDistance-ForestCONNECTSlopeGrasslandCOHESION+SlopeDevelopedCONNECTGrasslandCONNECT+DevelopedCONNECTGrasslandPAFRACSlope+SlopeDevelopedCONNECTWaterCONNECT+DevelopedCONNECT%GrasslandAgricultureCLUMPY-%Grassland%GrasslandAgriculturePAFRAC+%GrasslandDistanceDevelopedCONNECT-DevelopedCONNECTForestCONNECTWaterCONNECT+ForestCONNECT%AgricultureAgricultureCOHESION+AgricultureCOHESIONWaterCOHESIONWaterCLUMPY+WaterCLUMPY*thepredictorwiththelowestpartialcorrelationinthefullmodelwasremoved.
10TableS3.
Percentsignificant(P<0.
05)localr,rangeoflocalr,andmeanlocalrforfive,15and25nearestneighborsingroupsofwhite-taileddeerinnorthernIllinois(NIL),DuPageCounty(DuP),andWisconsin(WI).
GroupNumberofNearestNeighbors51525%P<0.
051MaxrMeanr%P<0.
051MaxrMeanr%P<0.
051MaxrMeanrAdultMales5.
70.
160.
134.
40.
110.
087.
90.
080.
06MaleYearlings7.
00.
280.
1711.
60.
180.
0914.
10.
120.
06MaleFawns9.
30.
190.
1511.
30.
090.
078.
20.
060.
05AdultMalesandFemaleYearlings6.
40.
270.
147.
60.
120.
088.
10.
090.
06AdultFemales14.
70.
320.
1618.
80.
240.
0920.
50.
150.
07FemaleYearlings5.
70.
160.
124.
80.
110.
074.
80.
070.
05FemaleFawns17.
10.
240.
1415.
20.
130.
0919.
50.
090.
06AdultFemalesandFawns16.
00.
310.
1622.
80.
230.
1024.
50.
190.
081NumberofautocorrelationcoefficientsthatweresignificantatP<0.
05dividedbythetotalnumberautocorrelationcoefficientscalculatedforeachgroup*100.
Includingonlysignificantlocalrvalues.
Kelly,NohraE.
Mateus-Pinilla,WilliamBrown,MarilynO.
Ruiz,MarlisR.
Douglas,MichaelE.
Douglas,PaulShelton,TomBeissel,JanNovakofskiMicrosatelliteMarkersThefollowingmicrosatelliteswereemployedinthisstudy:BM1225,BM4107,CSN3,(Bishopetal.
1994),IGF-1(Kirkpatrick1992),OBCAM(Friesetal.
1993),OarFcb304(Buchananetal.
1993),RT20,RT23,RT27(Wilsonetal.
1997)andSrcrsp-10(Bhebheetal.
1994).
Welabeledforwardprimerswithfluorescentdyes(NED,HEX,FAM)andseparatedmicrosatellitefragmentsonanABI3730XLcapillarysequencer(AppliedBiosystems,FosterCity,CA).
WevisualizedmicrosatellitegenotypeswithGeneMapper(v.
4.
0;AppliedBiosystems,FosterCity,CA).
WeusedMicro-checker(v.
2.
2.
3;VanOosterhoutetal.
2004)toevaluategenotypingerrorsusingexpectedallelefrequenciesderivedunderHardy-Weinbergequilibrium(HWE).
FSTSurfaceProjectionWeusedtheSingleSpeciesGeneticDivergenceoptionwithintheGeneticLandscapesGIS(GeographicInformationSystem)ToolboxtoprojectasurfacefrompairwiseFSTvaluescalculatedbetweenall31studysites.
TheprogramfirstassociatedpairwiseFST2valueswithmidpointsbetweenallstudysitesandanetworkofnearestneighbors.
Spatialinterpolationwasthenperformedusinganinversedistanceweightedinterpolationalgorithmtoestimategeneticdistancesalongagridoverlaidonthestudyarea.
GeneticdistancesforallpointsacrossthegridwereinterpolatedsuchthatmidpointFSTvaluesthatwerespatiallycloserinfluencedtheestimatemoresothanthosethatweredistant.
Moredetailsontheinterpolationprocedurearedescribedinhttp://www.
werc.
usgs.
gov/productdetails.
aspxid=4017.
FRAGSTATSmetricsTheConnectanceIndex(CONNECT)measuresfunctionalconnectivity,meaningthatgridcellsinthedatathatdepictthetargetvariablearenotliterallyadjacent,buttheyareconsideredadjacent(orconnected)withinagiventhresholddistance.
Inthiscase,adjacencywasdefinedascellswithin100mofeachother.
Theuser-defined100mthresholdwasusedtoaccountforpotentialimprecisionofdataclassificationsatfinespatialresolutionsandtoprovideamorerealistic(i.
e.
,functional)depictionofhowdeermightinteractwiththelandscape.
Themetricitselfisapercentage,witharangeof0to100.
Morespecifically,itmeasuresthepercentageoftargetvariableadjacencies(connectionsorjoins)relativetoallpossibleadjacencies.
FormoreinformationontheConnectanceIndexsee:http://www.
umass.
edu/landeco/research/fragstats/documents/Metrics/Connectivity%20Metrics/Metrics/C122%20-%20CONNECT.
htm3ThePatchCohesionIndex(COHESION)isasecondmeasureofconnectivityofalandscapevariable.
Thismetrictakesintoaccountphysicaladjacency(withoutathreshold)incombinationwiththesizeandshapeofthepatches.
Takingforestasanexample,ahigherCOHESIONvaluewouldoccurinalandscapewithlargerandcompactpatchescomparedtoonewithsmallorconvolutedpatches.
FormoreinformationonthePatchCohesionIndexsee:http://www.
umass.
edu/landeco/research/fragstats/documents/Metrics/Connectivity%20Metrics/Metrics/C121%20-%20COHESION.
htmTheClumpinessIndex(CLUMPY)isametricindicatinghowcontiguousordispersedaretheadjacentpatchesofalandscapevariable.
AhighervalueofCLUMPYwouldoccurifseveralpatcheswerelocatedclosetogetherratherthanbeingmoreuniformlydistributed.
FormoreinformationontheClumpinessIndex(CLUMPY)seehttp://www.
umass.
edu/landeco/research/fragstats/documents/Metrics/Contagion%20-%20Interspersion%20Metrics/Metrics/C115%20-%20CLUMPY.
htmThePerimeter-AreaFractalDimension(PAFRAC)isashapemetricdeterminedacrossarangeofspatialscales.
PARFRACislowforpatcheswithsimpleperimetersandincreasesforpatchshapeswithhighlyconvolutedperimeters.
FormoreinformationonthePerimeter-AreaFractalDimensionIndex(PAFRAC),seehttp://www.
umass.
edu/landeco/research/fragstats/documents/Metrics/Shape%20Metrics/Metrics/C23%20-%20PAFRAC.
htm.
Multivariatelinearregressionanalysis4DescriptionandsourceoflandscapevariablesincludedinmultivariateregressionanalysisarelistedinTableS1.
Topreventoverlyinfluentialobservationsfrombiasingourmodels,weusedleveragescores,Cook'sDvalues,andstandardizedinfluencevaluestoidentifyoutliers(Kieetal.
2002;ChatterjeeandHadi2009;Anlaufetal.
2011).
Leveragescoresidentifyobservationsthatresultinlargechangesinregressionlinefitupontheirdeletion.
Wecalculatedleverage(pi)accordingtoChatterjeeandHadi(1986)andconsideredobservationsoverlyinfluentialwhenpi>2p/N(p=numberofindependentvariablesinthemodel;N=numberofobservations).
Cook'sDvalueswerecalculatedaccordingtoCook(1977)andcomparedtoanFdistributionwithα=0.
05and(N-p)degreesoffreedom.
AllCook'sDvalues>thecriticalFvaluewereconsideredoverlyinfluentialandremovedfromthemodel(Cook1977).
LeveragescoresandCook'sDallowedustodeterminetheeffectsofoutliersontheoverallmodel,butstandardizedinfluencevalues(DFFITS)allowedustoexaminetheinfluenceofeachobservationonitspredictedvalue.
WecalculatedDFFITSaccordingtoChatterjeeandHadi(1986)andeliminatedobservationsyieldingvalues>2)/(Np(ChatterjeeandHadi1986).
Usingthesethreecriteria,weidentifiedthirteenobservationsoutof465(2.
8%)thatwereoutliersandafterstringentlyevaluatingtheirbasis(Motulsky2010),weomittedthemduringfurtheranalyses.
Themajorityoftheoutliersremoved(7/13)involvedstudysitesthathadrelativelylowsamplesizes.
Threeofthirteenoutliersinvolvedpairwisecomparisonswithstudysite27,thoughtheremainingtenoutliersappearedtoinvolvestudysitesthatwererandomlydistributedgeographically.
AsingleoutlierhadthehighestFSTvalueobserved,thoughtheremainingoutliersdidnotexhibitunusuallyhighorlowFSTvaluesascomparedtotherestofthe5dataset.
WecomparedvaluesofdependentvariablesofoutlierstovaluesfortherestofthedatabyexaminingboxplotsandplottingdependentvariablesagainstFSTvalues(datanotshown).
Trendsinthedistributionofvaluesfordependentvariablewerenotapparentinoutliersascomparedtotherestofthedata.
Whentwoormorelandscapevariableswerehighlycorrelated(Pearson'srP>0.
7),thepredictorwiththelowestpartialcorrelationinthefullmodelwasremoved.
RemovinglandscapevariableswithrP>0.
7(n=7)resultedinagenerallackofcollinearityamongpredictorsasdeterminedbyvarianceinflationfactors.
CorrelatedpredictorsthatwereremovedfromthemodelarelistedinTableS2.
Weusedvarianceinflationfactors(VIF)toevaluatetheincreaseinvarianceforestimatedregressioncoefficientsresultingfromcollinearpredictors,withVIF>10indicativeofhighmulticollinearity(Kutneretal.
2004).
Afterremovinghighlycorrelatedvariables,wecalculatedvarianceinflationfactorsforindependentvariablesandfoundthatthevarianceofestimatedregressioncoefficientswasnotsubstantiallyincreasedbycollinearpredictorsasVIFvaluesforallpredictorswere0.
7thatweresubsequentlyremovedfromthemodel.
VariableCorrelateDirectionofCorrelationVariableRemoved*%GrasslandSlope+%GrasslandForestCONNECTDevelopedCONNECT+ForestCONNECT%GrasslandGrasslandCONNECT-%GrasslandForestCONNECTGrasslandCONNECT+ForestCONNECTForestCONNECTWaterCONNECT+ForestCONNECTAgricultureCLUMPY%Agriculture-AgricultureCLUMPY%RiparianSlope+Slope%GrasslandForestCLUMPY-%GrasslandForestCONNECTDistance-ForestCONNECTSlopeGrasslandCOHESION+SlopeDevelopedCONNECTGrasslandCONNECT+DevelopedCONNECTGrasslandPAFRACSlope+SlopeDevelopedCONNECTWaterCONNECT+DevelopedCONNECT%GrasslandAgricultureCLUMPY-%Grassland%GrasslandAgriculturePAFRAC+%GrasslandDistanceDevelopedCONNECT-DevelopedCONNECTForestCONNECTWaterCONNECT+ForestCONNECT%AgricultureAgricultureCOHESION+AgricultureCOHESIONWaterCOHESIONWaterCLUMPY+WaterCLUMPY*thepredictorwiththelowestpartialcorrelationinthefullmodelwasremoved.
10TableS3.
Percentsignificant(P<0.
05)localr,rangeoflocalr,andmeanlocalrforfive,15and25nearestneighborsingroupsofwhite-taileddeerinnorthernIllinois(NIL),DuPageCounty(DuP),andWisconsin(WI).
GroupNumberofNearestNeighbors51525%P<0.
051MaxrMeanr%P<0.
051MaxrMeanr%P<0.
051MaxrMeanrAdultMales5.
70.
160.
134.
40.
110.
087.
90.
080.
06MaleYearlings7.
00.
280.
1711.
60.
180.
0914.
10.
120.
06MaleFawns9.
30.
190.
1511.
30.
090.
078.
20.
060.
05AdultMalesandFemaleYearlings6.
40.
270.
147.
60.
120.
088.
10.
090.
06AdultFemales14.
70.
320.
1618.
80.
240.
0920.
50.
150.
07FemaleYearlings5.
70.
160.
124.
80.
110.
074.
80.
070.
05FemaleFawns17.
10.
240.
1415.
20.
130.
0919.
50.
090.
06AdultFemalesandFawns16.
00.
310.
1622.
80.
230.
1024.
50.
190.
081NumberofautocorrelationcoefficientsthatweresignificantatP<0.
05dividedbythetotalnumberautocorrelationcoefficientscalculatedforeachgroup*100.
Includingonlysignificantlocalrvalues.
- connectivity网易轻博客相关文档
- 香港网易轻博客
- multiple网易轻博客
- 金融服务外包动态
- x2100网易轻博客
- 意象网易轻博客
- descriptive网易轻博客
Hostodo：4款便宜美国vps七折优惠低至$13/年;NVMe阵列1Gbps带宽,免费DirectAdmin授权
hostodo怎么样?快到了7月4日美国独立日，hostodo现在推出了VPS大促销活动，提供4款Hostodo美国独立日活动便宜VPS，相当于7折，低至$13/年，续费同价。Hostodo美国独立日活动结束时间不定，活动机售完即止。Hostodo商家支持加密数字货币、信用卡、PayPal、支付宝、银联等付款。Hostodo美国独立日活动VPS基于KVM虚拟，NVMe阵列，1Gbps带宽，自带一个...
瓜云互联-美国洛杉矶高防CN2高防云服务器，新老用户均可9折促销！低至32.4元/月！
瓜云互联一直主打超高性价比的海外vps产品，主要以美国cn2、香港cn2线路为主，100M以内高宽带，非常适合个人使用、企业等等！安全防护体系 弹性灵活，能为提供简单、 高效、智能、快速、低成本的云防护，帮助个人、企业从实现网络攻击防御，同时也承诺产品24H支持退换，不喜欢可以找客服退现,诚信自由交易！官方网站：点击访问瓜云互联官网活动方案：打折优惠策略：新老用户购买服务器统统9折优惠预存返款活动...
BuyVM迈阿密KVM上线,AMD Ryzen 3900X+NVMe硬盘$2/月起
BuyVM在昨天宣布上线了第四个数据中心产品：迈阿密，基于KVM架构的VPS主机，采用AMD Ryzen 3900X CPU，DDR4内存，NVMe硬盘，1Gbps带宽，不限制流量方式，最低$2/月起，支持Linux或者Windows操作系统。这是一家成立于2010年的国外主机商，提供基于KVM架构的VPS产品，数据中心除了新上的迈阿密外还包括美国拉斯维加斯、新泽西和卢森堡等，主机均为1Gbps带...
网易轻博客为你推荐
-
买虚拟主机虚拟主机购买要注意哪些？？？租用主机一般公司网站租用主机选哪一种的？linux主机linux主机与Windows主机的区别?谢谢中文域名注册查询域名还分中文和英文的吗，在哪里可以查到中文域名到期了？域名主机域名和主机名之间的区别是什么便宜的虚拟主机低价虚拟主机那种类型的好呢？jsp虚拟空间java虚拟主机空间怎么选择，国内jsp虚拟主机比较稳定java项目做好后需要推荐一下吧深圳网站空间怎么样建立网站青岛虚拟主机虚拟主机在什么地方买好？又便宜？二级域名什么是二级域名