第五章分子系統(tǒng)發(fā)育分析課件

第五章分子系統(tǒng)發(fā)育分析5.1分子進(jìn)化的基本概念5.2分子進(jìn)化模型與序列分歧度計算5.3分子系統(tǒng)樹的構(gòu)建5.4分子系統(tǒng)樹的檢驗5.5分子系統(tǒng)發(fā)育分析軟件及應(yīng)用5.1分子進(jìn)化的基本概念系統(tǒng)發(fā)生（phylogeny）——是指生物形成或進(jìn)化的歷史系統(tǒng)發(fā)生學(xué)(phylogenetics)——研究物種之間的進(jìn)化關(guān)系系統(tǒng)發(fā)生樹（phylogenetictree）——表示形式，描述物種之間進(jìn)化關(guān)系5.1分子進(jìn)化的基本概念同源性與相似性關(guān)于現(xiàn)代人起源的研究:

線粒體DNA ——所有現(xiàn)代人都是一個非洲女性的后代分類單元（物種或序列）物種之間的進(jìn)化關(guān)系

有根樹與無根樹有根樹的數(shù)目ABCABCDDDDD無根樹的數(shù)目ABCABCDDDNumberofTaxaNumberofunrootedtrees

Numberofrootedtrees

313431551510561059457945103958103951351359135135202702510202702534459425有根樹與無根樹的數(shù)目5.2分子進(jìn)化模型

與序列分歧度計算5.2.1核苷酸序列進(jìn)化5.2.2蛋白質(zhì)編碼序列進(jìn)化5.2.3核苷酸序列分歧度5.2.4蛋白質(zhì)編碼序列分歧度5.2.1核苷酸序列進(jìn)化ATCGJukesandCantor(1969)的單參數(shù)模型嘌呤嘧啶在t時間內(nèi)DNA序列上某個位點的堿基由A突變到G的概率為在t時間內(nèi)DNA序列上某個位點的堿基為A保持不變的概率為JukesandCantor(1969)的單參數(shù)模型在t+t時間內(nèi)核苷酸在某個位點上堿基為A保持不變的概率為表示在0到t這段時間內(nèi)DNA序列上某個位點的堿基為A保持不變的概率因此，在0到t這段時間內(nèi)DNA序列上某個位點的堿基保持不變的概率為因此，在0到t這段時間內(nèi)DNA序列上某個位點的堿基發(fā)生突變的概率為1下面考慮在0到t這段時間內(nèi)DNA序列上某個位點的堿基發(fā)生突變的平均次數(shù)設(shè)表示0到t這段時間內(nèi)DNA序列上某個位點的堿基發(fā)生突變的次數(shù)為n的概率，則突變率2即在0到t這段時間內(nèi)DNA序列上某個位點的堿基發(fā)生突變的次數(shù)服從泊松分布,平均次數(shù)為，方差也為。兩條DNA序列分歧度的計算AACGACGATCGAAGGACGATCG:Species2AACGATGATCG:Species1tt

Thetimeis2tbetweenSpecies1andSpecies2

定義兩條DNA序列間的分歧度為K=2t

對于JukesandCantor模型Sp1:AAGCCTCGGGGCCCTTATTTTTTG||||||||||||||||||Sp2:AATCTCCGGGGCCTCTATTTTTTTp=0.25K=0.304099Geneticdistancesarescaledtobethenumberofsubstitutionspersite.Kimura(1980)的兩參數(shù)模型ATCG嘌呤嘧啶

堿基的轉(zhuǎn)換(transition)

堿基的顛換(transversion)(1)(2)(3)(4)(1)－(3)代入(1)式，得推導(dǎo)KK80AACGACGATCGAAGGACGATCG:Species2AACGACTATCG:Species1tt

Thetimeis2tbetweenSpecies1andSpecies2

定義兩條DNA序列間的分歧度為K=2t=2(+2)t

設(shè)Ns為轉(zhuǎn)換數(shù)，Nv為顛換數(shù)，則Ps=P=Ns/L,Pv=Q=Nv/LSp1:AAGCCTCGGGGCCCTTATTTTTTG||||||||||||||||||Sp2:AATCTCCGGGGCCTCTATTTTTTTWhatarePandQ?P=4/24,Q=2/24K值的取樣誤差其中P=4/24Q=2/24p=6/24L=24Implicationsofsubstitutionsinprotein-codingregionsNonsynonymoussubstitution:ThrTyrLeuLeuACCTATTTGCTGACCTCTTTGCTGThrSer

LeuLeuSynonymoussubstitution:ThrTyrLeuLeuACCTATTTGCTGACCTCCTTGCTGThrTyrLeuLeuTheratesofnucleotidesubstitutionsinthethirdpositionaremuchhigherthaninthefirstandsecondpositions,duetoredundancyinthethirdposition:Changesinthefirstandsecondpositionfrequentlychangestheresultingaminoacid,whilechangesinthethirdpositionaretypicallysynonymous.Countingthenumberofsynonymous&nonsynonymoussubstitutionsbetweenapairofhomologoussequences:Inthe“evolutionarymethod,”wetakeintoaccountallpossibleevolutionarypathwaysbetweeneachpairofhomologouscodons.sdandndarethenumberofsynonymous&nonsynonymoussubstitutionspercodon.ForGTT(Val)andGTA(Val),thereisonesynonymousdifferenceandnononsynonymousdifferences.Therefore,sdandndare1&0,respectively.Itismorecomplicatedwhentherearemultiplesubstitutionspercodon:Multiplesubstitutionspercodon:Thereare6pathwaysbetweenTTGandAGA:TTG(Leu)<->ATG(Met)<->AGG(Arg)<->AGA(Arg)TTG(Leu)<->ATG(Met)<->ATA(Ile)<->AGA(Arg)TTG(Leu)<->TGG(Trp)<->AGG(Arg)<->AGA(Arg)TTG(Leu)<->TGG(Trp)<->TGA(Ter)<->AGA(Arg)TTG(Leu)<->TTA(Leu)<->ATA(Ile)<->AGA(Arg)TTG(Leu)<->TTA(Leu)<->TGA(Ter)<->AGA(Arg)Subs:S,N1,20,31,2*1,2*Wecanignore4&6,whichinvolvestopcodons.sdandndarethen?and9/4,respectively.ThetotalnumberofsubstitutionsThetotalnumberofsubstitutionsareSdandNd,whicharethesumsofsdandndforallcodonsinthecomparedsequences.NotethatSd+Ndisequaltothetotalnumberofnucleotidedifferencesbetweenthetwosequencescompared.Sincesomesubstitutionsmightbemorecommonduetotransition-transversionbias,wecanmodifyourestimateswithparametersfortheseratesTheproportionofsubstitutionsWhileSdandNd

provideuswithinformationaboutthenumberofsubstitutions,wearemoreinterestedintherelativeratesofthesesubstitutionsbetweendifferentgenes.Therefore,weestimatetheproportionofdifferences:Ks=Sd/S&KA=Nd/NWhereS&Narethenumbersofsynonymousandnonsynonymoussites;S+N=3C(thetotalnumberofcodons).TheratesKs&KAareestimatedusingtheJukes-Cantormethod,assumingequalnucleotidefrequencyandnotransition-transversionbias.Add’lparameterscanbeaddediftheseassumptionsarenotvalid.Seq1SerThrGluMetCysLeu

TCAACTGAGATGTGTTTASeq2TCAACAGAGATATGTCTASerThrGluIleCysLeu Sd=2

Nd=1Onesubstitutioninthecodon:Seq1SerThrGluMetCysLeu

TCGACAGAGATGTGTTTASeq2TCGACAGAGATGTGTCTTSerThrGluMetCysLeuSeveralpathwaysproblem1. TTACTA

CTT2. TTATTTCTTTwosubstitutioninthecodonSeveralpathwaysproblem

LeuLeuLeu1. TTA

CTA

CTT2. TTATTT

CTT

LeuPheLeu

Sd=2Nd=0

Sd=0Nd=2Averagepathways Sd=(2+0)/2=1

Nd=(0+2)/2=1Seq1 SerThrGluMetCysLeu

TCGACAGAGATGTGTTTASeq2 TCGACAGAGCGCTGTTTA SerThrGluArgCysLeuSeveralpathwaysproblem1. ATGATC

AGC

CGC2. ATGATC

CTC

CGC3. ATGAGGAGC

CGC4. ATGAGGCGGCGC5. ATGCTG

CTC

CGC6. ATGCTG

CGGCGCThreesubstitutioninthecodonSeveralpathwaysproblemMetIleSerArg1.ATGATC

AGCCGCMetIleLeuArg2.ATGATCCTCCGCMetArgSerArg3.ATGAGGAGCCGCMetArgArgArg4.ATGAGG

CGG

CGCMetLeuLeuArg5.ATG

CTG

CTC

CGCMetLeuArgArg6.ATG

CTG

CGG

CGC

Sd=0Nd=3AveragepathwaysSd=(0+0+0+2+1+1)/6=0.67Nd=(3+3+3+1+2+2)/6=2.33

Sd=0Nd=3

Sd=0Nd=3

Sd=2Nd=1

Sd=1Nd=2

Sd=1Nd=2Calculatethenumbersof

synonymousandnonsynonymoussites苯基丙氨酸半胱氨酸

SerThrGluMetCysLeuS1

TCAACTGAGATGTGTTTA NNSNNSNN1/3SNNNNN1/2S1/3SN1/3S

2/3N1/2N2/3N2/3N SerThrGluIleCysLeuS2 TCAACAGAGATATGTCTA NNSNNSNN1/3SNN2/3SNN1/2S1/3SNS

2/3N1/3N1/2N2/3NExample:Seq1SerThrGluMetCysLeu

TCAACTGAGATGTGTTTASeq2TCAACAGAGATATGTCTASerThrGluIleCysLeuS=4.1667N=13.8333Sd=0+1+0+0+0+1=2Nd=0+0+0+1+0+0=1DivergenceforSynonymousandNonsynonymoususingJukesandCantor’sModelinThisExample5.3分子系統(tǒng)樹的構(gòu)建5.3.1距離矩陣法5.3.2簡約法UPGMA法鄰接法Fitch-Margoliash法最大簡約法進(jìn)化簡約法其他方法方法比較

距離矩陣法—UPGMA法設(shè)類群OTUp和OTUq中含有np和nq個原始類群，并且它們聚合成類群OTUr，則OTUr與OTUi間的距離dr,i為例5—1OTU1OTU2OTU3OTU4

OTU1OTU2

OTU3

OTU4UnweightedPair-GroupMethodusingArithmeticaverages第一步：將OTU1和OTU2聚合為OTUr1,則OTUr1OTU3OTU4OTUr1OTU3OTU4第二步：將OTUr1和OTU3聚合為OTUr2,則OTUr2OTU4OTUr2

OTU4第三步：將OTUr2和OTU4聚合為OTUr3,結(jié)束。用UPGMA法構(gòu)建的系統(tǒng)樹

距離矩陣法—鄰近法ABCDEABCDEABCDEABCDEABCDE鄰近法的計算步驟對于所有的分類單元i，計算選擇一對分類單元i和j，使最小將i和j歸并為新的類(ij)，在樹中添加一個新的節(jié)點，代表新生成的分類，計算從i和j到新節(jié)點的分支長度：計算新類與其它類的距離：如果有兩個以上的分類存在，則繼續(xù)執(zhí)行循環(huán)；否則用長度為Di,j的分支連接剩余的兩個類。例5-25個分類群5SrRNA的例子BsuBstLviAmoMluBsuBstLviAmoMlu0.30930.3387670.3958670.4524670.420533設(shè)LviAmo0.11140.1681BsuBstLviAmoMluBsuBstLviAmoMluu1()()u10.12220.17980.17980.12220.27190.2719BsuBstu1

MluBsuBstu1

Mlu0.26310.27850.28690.3551BsuBstu1

Mlu設(shè)LviAmo0.11140.1681Bsu0.04920.0730BsuBstu1

MluBsuBstu1

Mlu

u2BstMlu

u2

BstMlu

u2BstMlu

u2

BstMlu0.30580.32040.3970

u2BstMluLviAmo0.11140.1681Bsu0.04920.0730設(shè)Bst0.06460.0500Mlu0.1412

距離矩陣法

—Fitch-Margoliash法距離定義：某個分類單元到一個合并類的距離定義為這個分類單元到這個合并類中的分類單元的平均距離ACED例5—3設(shè)A—D4個類群間的距離為：ABCDABCDACAve.BDACBDAC①②③ABCDABCDBD(AC)B

D(AC)ACBD①②③②－③：④①+④調(diào)整ACDB

最大簡約法推斷序列中堿基替換的數(shù)為最小的進(jìn)化樹適用于較短的、相似度較高的序列算法并不復(fù)雜，但能保證獲得最優(yōu)樹常用的軟件有PAUP和PHYLIP以一個例子說明最大簡約法位點123456789序列1AAGAGTGCA序列2AGCCGTGCG序列3AGATATCCA序列4AGAGATCCGAdaptedfromLiandGraur1991四條序列可能有3種無根樹一些位點有信息，這些點偏愛其中的一棵樹，信息位點的字母至少在兩條序列中出現(xiàn)只須分析信息位點，如位點5、7、9GCAA1234位點123456789序列1AAGAGTGCA序列2AGCCGTGCG序列3AGATATCCA序列4AGAGATCCGGGACA1324GACA1423樹1樹2樹3AAAAAACTG1234位點123456789序列1AAGAGTGCA序列2AGCCGTGCG序列3AGATATCCA序列4AGAGATCCGCATCG1324AGCT1423樹1樹2樹3TTCATGGAA1234位點123456789序列1AAGAGTGCA序列2AGCCGTGCG序列3AGATATCCA序列4AGAGATCCGGGAGA1324GAGA1423樹1樹2樹3AAAAAAGAG1234位點123456789序列1AAGAGTGCA序列2AGCCGTGCG序列3AGATATCCA序列4AGAGATCCGAAAGG1324AGGA1423樹1樹2樹3AAGAAGGAA1234GGAGA1324GAGA1423AAAAAGGCC1234GGCGC1324GCGC1423CCCCCAGAG1234AAAGG1324AGGA1423AAGAA位點5位點7位點9堿基替換總數(shù)455GGAGGGACAACGGGAACA位點123456789序列1AAGAGTGCA序列2AGCCGTGCG序列3AGATATCCA序列4AGAGATCCG

進(jìn)化簡約法以四個分類群為例說明考慮三種可能的系統(tǒng)樹，分別稱為X型、Y型、Z型：ABCDACBDADBC位點1234567891011121314151617181920SAAUCAGGCUUGCACUAACUGGSBAGGAGAAGUAAGGCCACUUCSCAGGUGUAAUCAGGGCAGAACSDAGCUACCUUUUGCAACGAUA將每個位點的堿基轉(zhuǎn)換成單位矢量11111333133111331112123413311341111