Evolutionary relationships between gene

1

'R÷UXVDO2OPD\DQ*|PPH7HNQLNOHUL$OWÕQGD*HQ'L]LOHULQLQ(YULPVHO

øOLúNLOHUL

Evolutionary Relationships Between Gene Sequences via Nonlinear

Embedding

7XQFD'R÷DQ

1

_{%LOJH.DUDoDOÕ}

2

1.

%L\RWHNQRORMLYH%L\RPKHQGLVOLN

%|OP

,

ø]PLU<NVHN7HNQRORML(QVWLWV

tuncadogan@iyte.edu.tr

2.

(OHNWULNYH(OHNWURQLN0KHQGLVOL÷L%|OP

ø]PLU<NVHN7HNQRORML(QVWLWV

bilgekaracali@iyte.edu.tr

g]HWoH

%XoDOÕúPDGDoRNOXGL]LKL]DODPDVÕQÕWDNLEHQoÕNDUÕPVDQPÕú JHQ GL]L LNLOLOHUL DUDVÕQGDNL HYULPVHO X]DNOÕNODUD uygulanan GR÷UXVDOROPD\DQJ|PPHWHNQL÷LQLQKDWDDQDOL]LVXQXOPXúWXU %X DPDoOD VHQWHWLN JHQ GL]LOHUL ROXúWXUXOPXú YH o D\UÕ HYULPVHO \RO ER\XQFD KHU GL]L LNLOLVL DUDVÕQGDNL JHUoHN HYULPVHO X]DNOÕN ND\ÕW DOWÕQD DOÕQDUDNWDQ UDVWODQWÕVDO yer GH÷LúWLUPH PXWDV\RQXQD WDEL WXWXOPXúODUGÕU 'úN ER\XWOX ELU YHNW|U X]D\ÕQD \DSÕODQ GR÷UXVDO ROPD\DQ J|PPH LúOHPL |QFHVL YH VRQUDVÕ HOGH EXOXQDQ oÕNDUÕPVDQPÕú HYULPVHO X]DNOÕNODULOHJHUoHNHYULPVHOX]DNOÕNODUDUDVÕQGDNLIDUNOÕOÕNODU NDUúÕODúWÕUÕOPÕúWÕU 6RQXoODU GR÷UXVDO ROPD\DQ J|PPH LúOHPLQLQ oÕNDUÕPVDQDQ HYULPVHO X]DNOÕNODUGD \HU DODQ KDWD SD\ÕQÕ GúUPH NRQXVXQGDNL EDúDUÕVÕQÕ J|VWHUPLúWLU 6RQXo olarak, HYULPVHO X]DNOÕNODU ]HULQH X\JXODQDQ GR÷UXVDO ROPD\DQ J|PPH LúOHPLQLQ JHQ GL]LOHUL DUDVÕQGDNL HYULPVHO LOLúNLOHUL RUWD\D oÕNDUPDN DGÕQD JYHQLOLU oÕNDUÕPODUGD EXOXQDELOHFH÷LEHOLUWLOPLúWLU

Abstract

We present an error analysis on the application of non-linear embedding on pairwise evolutionary distances inferred over a collection of genetic sequences following multiple sequence alignment. To this end, we have generated gene sequences evolved by random substitutions along three different evolutionary pathways with known evolutionary distances between every sequence pair. We have compared the discrepancy between the inferred evolutionary distances to the true distances before and after non-linear embedding into a low dimensional vector space. The results indicate that non-linear embedding achieves significant reduction in error in the estimated evolutionary distances. Consequently, non-linear embedding of evolutionary distances can provide more reliable inferences on the evolutionary relationships between genetic sequences.

1.

*LULú

%L\RORMLN RUJDQL]PDODUÕQ JHQ YH SURWHLQ GL]LOHUL DUDVÕQGDNL HYULPVHO LOLúNLOHULQ NHúfi, sistemlerini meydana getiren

PROHNOHU YH IL]\RORMLN PHNDQL]PDODUÕ RUWD\D oÕNDUPDN DoÕVÕQGDQ|QHPWDúÕU. %XoDOÕúPDODUHOGHNLGL]LOHULQDUDVÕQGDNL benzerlikleri dikkate alarak RUWDN DWDODUÕQ RUWD\D oÕNDUÕOPDVÕ

YH WP EX GL]LOHULQ DWDODU-WRUXQODU úHNOLnde birbirleri ile

LOLúNLOHQGLULOGLNOHUL VR\ D÷DoODUÕQÕQ ROXúWXUXOPDVÕ\OD JHUoHNOHúWLULOLU%LUGL÷HUGH\LúOHILORJHQHWLN\|QWHPOHUEXJHQ YHSURWHLQGL]LOHULDUDVÕQGDNLHQPXKWHPHOKL]DODPD\ÕGLNNDWH DODUDN PROHNOHU \HU GH÷LúLPOHULQL VDSWDPD\Õ YH EX úHNilde

GL]LOHULQ RUWDN JHoPLúlerLQL RUWD\D oÕNDUPD\Õ KHGHIOHU Bu

\|QWHPOHU GL]L X]DNOÕNODUÕQÕQ KHVDSODQÕS EX X]DNOÕNODUÕ HYULPVHO X]DNOÕNODUOD LOLúNLOHQGLUHQ matematiksel bir modelin

X\JXODQPDVÕYHVRQXoWDHYULPVHOX]DNOÕNODUÕQoÕNDUVDQPDVÕQD GD\DQÕU

PrDWLNWH JHQ YH SURWHLQ GL]LOHULQLQ oRN X]XQ ROPDVÕ YH

|]HOOLNOH JHoPLú ]DPDQD DLW GL]LOHULQ oR÷XQOXNOD HOGH EXOXQPDPDVÕVLVWHPLQNDUPDúÕNOÕ÷ÕQÕDUWWÕUÕUYHNDYUDQPDVÕQÕ JoOHúWLULU%XQHGHQGHQGROD\ÕEXWLSDUDúWÕUPDODULVWDWLVWLNVHO DQODPGD JHUoHNOHúWLULOLU %X X\JXODPDODU JHUHN GR÷UXOXN JHUHNVH KÕ] JLEL |]HOOLNOHUL GROD\ÕVÕ LOH LúOHPVHO RUWDPGD \UWOUKonvansiyonel DQODPGDNXOODQÕODQEX\|QWHPOHUGHQ

ED]ÕODUÕ -XNHV-Cantor mRGHOL >@ .LPXUD¶QÕQ iki parametreli modeli [2] ve genel zaman tersinir model¶GLU [3]. Bu klasik

\|QWHPOHU ELOLQPH\HQ HYULPVHO X]DNOÕNODU KDNNÕQGD ED]Õ

tahminlerde bulunsa da, bu tahminler bazen VRQXoODUÕQ

JYHQHELOLUOL÷LQLGúUHQKDWDODUD tabi olabilmektedir.

%X WLS oDOÕúPDODUGD VD÷OÕNOÕ VRQXoODU HOGH HGLOPHVLQLQ ve

EX VRQXoODUÕQ NDYUDQDELOLU J|UVHO LIDGHOHUOH DoÕNODQPDVÕQÕQ NDUúÕVÕQGDNL HQ E\N HQJHOOHUGHQ ELULVL de oRN ER\XWOXOXN

VRUXQXGXU $QDOL]H JLUGL RODUDN YHULOHQ EX GL]LOHULQ ]HULQGH NRGODQPÕú KHU ELU NDUDNWHU DQDOL]H NDWÕOPÕú ELU \HQL ER\XWX

ifade eder ve bu girdiler kimi zaman onbinlerce karakter

X]XQOX÷XQGD ROGX÷XQGD JHQ YH SURWHLQ GL]LOHULQGH ROGX÷X

gibi) boyut sorunu oDOÕúPDGDQ VD÷OÕNOÕ VRQXoODU DOPD\Õ KD\OL

JoOHúWLUHELOLU. %X VRUXQX DúPDN LoLQ GR÷UXVDO YH GR÷UXVDO

ROPD\DQ J|PPH WHNQLNOHUL NXOODQÕOPDNWDGÕU. Bu teknikler

2

PDNXO ER\XWODUGD LIDGH HWPHN LoLQ X\JXODQPÕúWÕU .ODVLN oRN ER\XWOX |OoHNOendirme [4@ YH HVDV ELOHúHQ Dnalizi [5] bu

WHNQLNOHULQ LON YH HQ VÕN NXOODQÕODQ |UQHNOHULGLU Analiz edilecek veULOHUELUG]OHP]HULQGHGD÷ÕOÕPJ|VWHUGL÷L]DPDQ

EDúDUÕOÕ VRQXoODU UHWPHVLQH UD÷PHQ EX \|QWHPOHULQ H÷UL \]H\OHUGH EDúDUÕVÕ] ROGX÷X OLWHUDWUGH EHOLUWLOPLúWLU >6@ (÷UL

\]H\OHUGHNL GD÷ÕOÕPODUÕ \DNDOD\DELOPHN LoLQ, bu klasik

\|QWHPOHULQX\DUODQPDVÕ ile ELUoRNGR÷UXVDOROPD\DQJ|PPH

WHNQL÷L JHOLúWLULOPLú YH IDUNOÕ DODQODUGD X\JXODQPÕúWÕU Bu

\|QWHPOHUGHQED]ÕODUÕGR÷UXVDOROPD\DQKaritalama [7], yerel

GR÷UXVDO J|PPH >8@ YH VWRNDVWLN \DNÕQOÕN J|PPHsi [9]

\|QWHPOHULGLU

'R÷UXVDO ROPD\DQ J|PPH \|QWHPOHULnin protein ve gen

GL]LOHUL DUDVÕQGDNL EHQ]HUOLNOHUL RUWD\D oÕNDUPDN DGÕQD oRNOX

dizi hL]DODPDVÕQÕ WDNLEHQ GL]L YHULOHULQGHQ oÕNDUÕPVDQDQ evULPVHO X]DNOÕNODUD X\JXODQPDVÕ GDKD |QFH oDOÕúÕOPÕúWÕU

gUQH÷LQ ELU oDOÕúPDGD VWRNDVWLN \DNÕQOÕN J|PPHsi WHNQL÷L kulODQÕODUDNELUELULLOH\DNÕQLOLúNLOHULoLQGHRODQED]ÕSURWHLQ dizileri DQODPOÕNPHOHULoLQGHJUXSODQPD\DoDOÕúÕOPÕúWÕU>0]. FDUNOÕ DQDOL] SDUDPHWUHOHULQLQ GHQHQLS EX SDUDPHWUHOHULQ

RSWLPL]H HGLOPH\H oDOÕúÕOGÕ÷Õ DUDúWÕUPDGD ED]Õ SURWHLQ

ailelerinin fonksiyonel oODUDN ELUELULQH \DNÕQ \HOHULQLQ evrimsel anlamda uzak SURWHLQOHUGHQ D\UÕOÕS iki boyutlu uzayda gruplaQPDVÕ EDúDUÕOPÕú EX úHNLOGH GR÷UXVDO ROPD\DQ

J|PPHWHNQLNOHULQLQSURWHLQYHJHQGL]LOHULQLNPHOH\HELOPH

becerisi ortaya koyulPXúWXU>0].

ø]RPHWULN|]HOOLNKaritalama (ISOMAP)GR÷UXVDOROPD\DQ

J|PPH \|QWHPOHULQGHQ ELULGLU %X \|QWHP NODVLN |OoHNOHQGLUPH LoLQH J|POPú RODUDN, NRPúXOXN GL\DJUDPÕ

]HULQGHNL MHRGH]LN X]DNOÕNODUÕ GLNNDWH DOÕU[6@ %X |]HOOLN

\|QWHPH H÷UL \]H\OHU ]HULQH GD÷ÕOPÕú YHULOHUL ELOH EDúDUÕOÕ úHNLOGH \DNDODPD EHFHULVLQL ND]DQGÕUÕU $QDOL]H JLUGL RODUDN YHULOHQ oÕNDUÕPVDQPÕú HYULPVHO X]DNOÕNODU NRPúX \DNÕQ QRNWDODUDUDVÕQGDNLMHRGH]LNX]DNOÕNODUÕKHVDSODPDNLoLQiyi bir

\DNODúÕP RODUDN \HU DOÕU >6@ 8]DN QRNWDODU LoLQVH EX NRPúX noktalar DUDVÕQGDQ ELUoRN NoN DWODPD VW VWH HNOHQHUHN

MHRGH]LNX]DNOÕNODUKHVDSODQDELOLU

%XELOGLULGHVHQWHWLNJHQGL]LOHULQLQ,620$3DOJRULWPDVÕ NXOODQÕODUDN X\JXODQPÕú RODQ GR÷UXVDO ROPD\DQ J|PPH VRQUDVÕ HOGH HGLOHQ HYULPVHO X]DNOÕNODUÕQÕQ KDWD DQDOL]L VXQXOPXúWXU %X UDSRUXQ VÕUDGDNL NÕVPÕQGD IDUNOÕ HYULPVHO \ROODU L]OH\HQ VHQWHWLN GL]LOHULQ UHWLOPHVL YH KDWD DQDOL]LQLQ JHUoHNOHúWLULOPHVL LoLQ NXOODQÕODQ SDUDPHWUH DoÕNODQPÕú WDNLS HGHQE|OPGHLVHVRQXoODURUWD\DNRQXOPXúYHEXVRQXoODUÕQ \RUXPODQPDVÕJHUoHNOHúWLULOPLúWLU

2.

<|QWHP

dDOÕúPDGDLON|QFHLúOHPVHORUWDPGDKD]ÕUODQPÕúRODQVHQWHWLN JHQGL]LOHULQLQDUDODUÕQGDNLHYULPVHOX]DNOÕNOar, konvansiyonel Jukes-Cantor mRGHOL >@ NXOODQÕODUDN oÕNDUÕPVDQPÕúWÕU 'DKD sonra bu YHULOHU ]Hrinde ,620$3 \|QWHPL uygulanarak,

GR÷UXVDO ROPD\DQ J|PPH VRQUDVÕ HYULPVHO X]DNOÕN GH÷HUOHUL HOGH HGLOPLúWLU. Bu iki grup HYULPVHO X]DNOÕN GH÷Hri, sentetik

GL]LOHUDUDVÕQGDNLJHUoHNHYULPVHOPHVDIHOHUOHNDUúÕODúWÕUÕODUDN KDWD DQDOL]L RUWD\D NR\XOPXúWXU +DWD PLNWDUODUÕ JUDILNOHU ]HULQGH NDUúÕODúWÕUÕODUDN ,620$3 \|QWHPL LOH \DSÕODQ GR÷UXVDO ROPD\DQ J|PPH LúOHPLQLQ oÕNDUÕPVDQPÕú HYULPVHO X]DNOÕNODU LoLQGH \HU DODQ KDWD PLNWDUODUÕQÕ DQODPOÕ RODUDN GúUPH\L EDúDUÕS EDúDUDPDGÕ÷Õ GL÷HU ELU GH\LúOH KDWD GúUPe kapasitesi LQFHOHQPLúWLU Sentetik gen dizilerinin

ROXúWXUXOPDVÕ GDKLO WP LúOHPOHU YH DQDOL]OHU 0$7/$%Š \D]ÕOÕPÕNXOODQÕODUDN\DSÕOPÕúWÕU

2.1. 6HQWHWLNJHQGL]LOHULQLQROXúWXUXOPDVÕ

6HQWHWLN JHQ GL]LOHULQLQ KD]ÕUODQPDVÕQGD o D\UÕ HYULPVHO \RO L]OHQPLú DQDlizler de EX o \RO LoLQ D\UÕ D\UÕ

JHUoHNOHúWLULOPLúWLU (YULPVHO \ROODU RULMLQDO VHQWHWLN JHQ GL]LVLQLQED]ÕE|OJHOHULQGHUDVWODQWÕVDOED]DO\HUGH÷LúWLUPHOHU \HUOHúWLULOPHVL LOH \HQL GL]LOHU ROXúWXUXOPDVÕQD GD\DQÕU Bu

RSHUDV\RQLúOHPVHORUWDPGDVLPXOHHGLOPLúWLU

+HU ELU GL]L NHQGLVLQGHQ ELU |QFH JHOHQ GL]LQLQ |QFHNL QHVOLQ HQ \DNÕQ DNUDEDVÕGÕU %X úHNLOGH QHVLOOHU LOHUOHGLNoH ROXúDQ\HQLGL]LOHURULMLQDO(ilk) GL]LGHQJLWWLNoHIDUNOÕODúÕUODU Her bir evrimsel yolda 501¶HU (orijinal diziyle beraber) dizi bulunur. +HU ELU GL]L QNOHRWLGGHQ ROXúXU $QDOL]GH

NXOODQÕODQ \HU GH÷LúWLUPH KÕ]Õ ¶GLU 'L÷HU ELU GH\LúOH YDU RODQELUGL]LGHQ\HQLELUGL]LROXúWXUXOXUNHQGL]Llerde yer alan

YHKHUELULQLELUQNOHRWLGLQNDSODGÕ÷ÕE|OJHGHQKHUELUL

ioLQeski dizideki QNOHRWLGLQD\QHQ\HQLGL]L\H\HUOHúWLULOPH ihtimali UDVWJHOH RODUDN IDUNOÕ QNOHRWLGGHQ ELULQLQ

\HUOHúWLULOPH LKWLPDOL LVH ¶GLU Evrimsel yol 1 (EY1), Evrimsel yol 2 (EY2) ve Evrimsel yol 3¶Q (EY3) olXúWXUXOPDVÕQGDL]OHQHQ\ROlar ùHNLO¶GHYHULOPLúWLU

HDWD GH÷HUOHUL (< (< YH (< LoLQ D\UÕ D\UÕ HOGH

HGLOPLúWLU %XQXQOD EHUDEHU KDWD DQDOL]L KHU ELU HYULPVHO \RO LoLQ GL]L LoLQGH \HU DODQ WP GL]L LNLOLOHUL LoLQ WRSOX úHNLOGH GH÷LO DUDODUÕQGD EHOLUOL QHVLO VD\ÕODUÕ EXOXnan diziler

LoLQD\UÕD\UÕRUWD\DNRQPXúWXU(10, 20, 50, 100 ve 200 nesil

IDUNOÕOÕNODU LoLQ %XQXQ QHGHQL DUDGDNL QHVLO VD\ÕVÕQÕQ GH÷LúLPL\OH EHUDEHU KDWD oranÕQÕQ GD úLGGHWOL úHNLOGH

GH÷LúPHVidir.

2.2. *HQ GL]LOHUL DUDVÕQGDNL oÕNDUÕPVDQPÕú evrimsel

X]DNOÕNODUÕQKHVDSODQPDVÕ

6HQWHWLN JHQ GL]LOHUL DUDVÕQGD \HU DODQ HYULPVHO X]DNOÕNODUÕQ oÕNDUÕPVDQPDVÕQGD NXOODQÕODQ NRQYDQVL\RQHO -XNHV-Cantor

PRGHOL VWRNDVWLN ELU \DNODúÕPGÕU %X YH EHQ]HUL

konvansiyonel modeller, LNLGL]LDUDVÕQGDNLGL]L uzaklÕ÷ÕQÕED]Õ parametreler ve basit bir formulasyon \DUGÕPÕile yine bu iki

GL]L DUDVÕQGDNL HYULPVHO X]DNOÕ÷D oHYLULUler. Jukes-Cantor

PRGHOLQNOHRWLG\HUGH÷LúWLUPHKÕ]ÕQÕoHúLWQNOHRWLGLQWP IDUNOÕ LNLOL NRPELQDV\RQODUÕ LoLQ HúLW NDEXO HGHU %XQXQ \DQÕQGDQNOHRWLGIUHNDQVODUÕYHGL]L]HULQGHNLE|OJHOHUDUDVÕ \HU GH÷LúWLUPH KÕ]ODUÕ GD HúLW NDEXO HGLOPLúWLU -XNHV-Cantor

PRGHOLQLQIRPXODV\RQXQXPDUDOÕGHQNOHPGHYHULOPLúWLU X]DNOÕ÷ÕQÕ WHPVLO HGHU dRNOX GL]L KL]DODPDVÕ VRQXFX RUWD\D oÕNDQGL]LOHU]HULQGHKHUELUE|OJHLoLQD\UÕD\UÕ\DSÕODQLNLOL NDUúÕODúWÕUPDODU VRQUDVÕQGD LNLOLdizi X]DNOÕNODUÕ RUWD\D oÕNDU

%X KHVDSODPDGD KHUKDQJL ELU E|OJHGH LNL GL]LGH GH D\QÕ QNOHRWLG EXOXQX\RUVD ELU HúOHúPH IDUNOÕ QNOHRWLGOHU YDUVD HúOHúPHPH ND\ÕW HGLOLU øúOHP VRQXQGD WRSODP HúOHúPHPH VD\ÕVÕ LNL GL]LGH GH ERúOXN ROPD\DQ WRSODP E|OJH VD\ÕVÕQD RUDQODQÕU'HQNOHPLQGHULYDV\RQXUHIHUDQVE|OPQGHYHULOHQ

3 2.3. *HQ GL]LOHUL DUDVÕQGDNL oÕNDUÕPVDQPÕú evrimsel

X]DNOÕNODUÕQ,620$3\|QWHPLLOHLúOHQPHVL

,620$3 \|QWHPL NHQGLVLQH JLUGL RODUDN YHULOHQ ELU X]DNOÕN PDWULVLQLDOÕSPDWULVWHDUDODUÕQGDNLPHVDIHOHUYHULOHQQRNWDODUÕ oRNER\XWOXYHNW|UX]D\ÕQGDVDELWQRNWDODUDDWDU%X úHNLOGH QPHULNX]DNOÕNPDWULVOHULQLGHJHRPHWULNRODUDNWHPVLOHWPLú

olur. <|QWHPGHQ DOÕQDQ oÕNWÕGD YHULQLQ J|POHFH÷L YHNW|U

X]D\Õ ER\XW VD\ÕVÕ LVWH÷H J|UH VHoLOLU *HQHOGH LúOHP ELUoRN IDUNOÕ ER\XW VD\ÕVÕ LoLQ DUGÕ VÕUD WHNUDUODQÕU YH LúOHP VRQXQGD DUDúWÕUPDFÕQÕQ EX ER\XW VD\ÕODUÕQGDQ ELULQL VHoLS GR÷UXVDO ROPD\DQJ|PPHQLQEXER\XWWDNLELUYHNW|UX]D\ÕQD\DSÕOPDVÕ VD÷ODQÕU %R\XW VD\ÕVÕQÕQ VD÷OÕNOÕ RODUDN VHoLOHELOPHVL LoLQ ER\XWVD\ÕVÕQDNDUúÕOÕNDUWÕNYDU\DQV GH÷HUOHULQLQ oL]LOGL÷L ELU

grafik verilir DUWÕN YDU\DQV JUDIL÷L. Bu grafikte, artan boyut

VD\ÕVÕ LOH EHUDEHU GúHQ DUWÕN YDU\DQV GH÷HULQLQ LQFHOHQPHVL YDU\DQVWDNLGúúQVRQDHULSPLQLPXPD\DNODúWÕ÷ÕHQGúN ER\XWVD\ÕVÕQÕQVHoLPLQLVD÷ODU0PNQROGX÷XNDGDUGúN ER\XWWDNL LIDGHQLQ VHoLOPHVLQLQ QHGHQL YHUL VHWLQLQ LoLQGHNL \DSÕ\ÕER]PDGDQ, VLVWHPLPDNVLPXPGHUHFHGHEDVLWOHúWLUPHN

KDWWDLPNDQYDUVDJ|UVHOJHRPHWULNELUWHPVLOLVD÷OD\DELOHFHN RODQYH\DER\XWWDoDOÕúPDNWDGÕU

%LU|QFHNLDGÕPGD-XNHV-&DQWRUPRGHOLLOHoÕNDUÕPVDQPÕú olan HYULPVHO X]DNOÕNODU ,620$3 \|QWHPLQH JLUGL RODUDN

YHULOHUHNGR÷UXVDOROPD\DQJ|PPHWHNQL÷LLOHoRNER\XWOXELU YHNW|U X]D\ÕQGD LIDGH HGLOPHOHUL VD÷ODQÕU ,620$3 \|QWHPL KHUJHQGL]LVLQLGL]LOHUDUDVÕLNLOLHYULPVHOX]DNOÕNODUÕGLNNDWH DODUDN oRN ER\XWOX X]D\GD \HU DODQ NRRUGLQDWODUÕ EHOLUOL ELU

nokta ile ifade eder. HHU ELU JHQ GL]LVLQH WHNDEO HGHQbu

QRNWDODU DUDVÕQGDNL PHWULN PHVDIHOHU, \|QWHPLQ VRQXo RODUDN

YHUGL÷LHYULPVHOX]DNOÕNODURODUDNND\GHdilir.

ùHNLO1: EYYH¶QROXúWXUXOPDVÕQGDL]OHQHQ\ROODU

3.

6RQXoODU

YHWDUWÕúPD

+DWD GúUPH NDSDVLWHVLQLQ RUWD\D NRQPDVÕ LoLQ NXOODQÕODQ SDUDPHWUH RUWDODPD PXWODN KDWD GH÷HUOHULGLU %X SDUDPHWUH

bir \|QWHPGHQ oÕNWÕ RODUDN DOÕQDQ bir dizi ikiliVL DUDVÕQGDNL

oÕNDUÕPVDQPÕú HYULPVHO X]DNOÕN GH÷HUL LOH EX LNL GL]L DUDVÕQGDNLJHUoHNHYULPVHOX]DNOÕ÷ÕQELUELUOHULQGHQoÕNDUÕOPDVÕ

LOH RUWD\D oÕNDQ IDUNÕQ PXWODN GH÷HU LoLQH DOÕQPDVÕ D\QÕ LúOHPLQ WP GL]L LNLOLOHUL LoLQ WHNUDUODQPDVÕ YH VRQXQGD HOGH

edilHQ WP GH÷HUlerLQ RUWDODPDVÕQÕQ DOÕQPDVÕ\OD HOGH HGLOLU

6RQXoWDELULNRQYDQVL\RQHO\|QWHPHYHGL÷HULGH,620$3¶H DLWROPDN]HUHLNLKDWDGH÷HULROXúXU

,620$3 LúOHPL VRQXQGD GR÷UXVDO ROPD\DQ J|PPH LúOHPLQLQ \DSÕODFD÷Õ YHNW|U X]D\ÕQÕQ ER\XW VD\ÕVÕQÕ VHoPek

LoLQ DUWÕN YDU\DQV JUDILNOHUL ùHNLO LQFHOHQPLúWLU $UWÕN YDU\DQVWDNL GúúQ VRQD HUGL÷L ER\XWODU RODQ (< YH (< LoLQER\XWOX(<LoLQVHER\XWOXYHNW|UX]D\ODUVHoLOPLúWLU

6RQXoODUÕQ LQFHOHQPHVLQLQ YH ELUELUOHUL LOH NDUúÕODúWÕUÕOPDVÕQÕQ J|UVHO DQODPGD YH NROD\ DQODúÕODELOLU ROPDVÕ LoLQ JUDILN J|VWHULPOHU KD]ÕUODQPÕú YH EX E|OPQ GHYDPÕQGDVXQXOPXúWXU%XNÕVÕPGDYHULOHQKHUELUúHNLOD\UÕ ELU HYULPVHO \RO (< (< YH (< LoLQ \DSÕOPÕú DQDOL]L WHPVLOHGHU+HUúHNLOGH[-ekseninde nesil uzDNOÕNODUÕ

YH \ HNVHQLQGH LVH LOJLOL QHVLO X]DNOÕNODUÕQD LOLúNLQ RUWD\D oÕNDQ KDWD GH÷HUL NRQYDQVL\RQHO -XNHV-Cantor

PRGHOLYH,620$3\|QWHPLLoLQLNLD\UÕEDUYHULOPLúWLU

4

ùHNLl 3, 4 ve 5¶GH J|UOG÷ ]HUH WP DQDOL]H WDEL

WXWXOPXú HYULPVHO \ROODU LoLQ QHVLO IDUND NDGDU ± 50

DUDVÕNRQYDQVL\RQHOPRGHOYH,620$3oÕNWÕODUÕELUELUOHULQH oRN \DNÕQ KDWD GH÷HUOHUL RUWD\D NR\PXúWXU 6RQXo RODUDN ,620$3 \|QWHPL 50 nesile kadar olan mesafelerde,

NRQYDQVL\RQHO \|QWHPLQ oÕNDUÕPVDGÕ÷Õ HYULPVHO X]DNOÕNODUGD \HUDODQKDWD\Õbelirgin RODUDN GúUPHPLúWLU %XQXQ QHGHQL

LVH EX DQDOL] úDUWODUÕQGD QHVLOH NDGDU RODQ PHVDIHOHUGH NRQYDQVL\RQHO PRGHOLQ ]DWHQ ROGXNoD GúN YH NDEXO

edileELOLU|OoGHKDWDlar YHUPLúROPDVÕGÕU

<LQH D\QÕ úHNLOOHUGH J|UOPHNWHGLU NL YH QHVLO PHVDIHOHUGH NRQYDQVL\RQHO PRGHOLQ UHWWL÷L KDWD ELU DQGD \NVHOPHNWHGLU %XQXQ \DQÕQGD ,620$3 VRQXQGD oÕNDQ GH÷HUOHU LoHULVLQGH NDODQ KDWD oRN GúN PHVDIHOHUGH J|]OHQHQ KDWD LOH \DNÕQ GH÷HUOHUH VDKLSWLU %X da, ISOMAP

LoLQ100 ve 200 nesil VÕUDVÕyla 0,65 ve 0,75 GH÷HUOHULQGHdizi

X]DNOÕNODUÕQD YH YH GH÷HUOHULQGH oÕNDUÕPVDQPÕú HYULPVHO X]DNOÕNODUD NDUúÕOÕN JHOLU PHVDIHOHUGH RUWD\D oÕNDQ oRNEHOLUJLQELUKDWDLQGLUJHPHNDSDVLWHVLQLLúDUHWHWPHNWHGLU

,620$3 \|QWHPLQLQ \NVHN QHVLO PHVDIHOHULQGHNL KDWDODUÕ LQGLUJHPHGH NL EX EDúDUÕVÕ DOJRULWPDQÕQ ELUELULQGHQ oRN X]DNWD \HU DODQ LNL QRNWD DUDVÕQGDNL PHWULN PHVDIH\L

hesaplarken, girdi olarak kendisine verileQ GH÷HUL GLNNDWH

DOPDN \HULQH ELUELULQH oRN \DNÕQ PHVDIHGHNL QRNWDODU DUDVÕQGD NoN DWODPDODU \DSDUDN EX LNL X]DN QRNWDQÕQ ELULQGHQ GL÷HULQH XODúPDVÕQD YH ROXúDQ EX PHVDIH\L GLNNDWH DOPDVÕQDGD\DQÕU

ùHNLO3: (< LoLQ NRQYDQVL\RQHO PRGHO -XNHV-Cantor) ile

,620$3DUDVÕQGDRUWDODPDPXWODNKDWDNDUúÕODúWÕUPDVÕ

ùHNLO4: (< LoLQ NRQYDQVL\RQHO PRGHO -XNHV-Cantor) ile

,620$3DUDVÕQGDRUWDODPDPXWODNKDWDNDUúÕODúWÕUPDVÕ

ùHNLO5: (< LoLQ NRQYDQVL\RQHO PRGHO -XNHV-Cantor) ile

,620$3DUDVÕQGDRUWDODPDPXWODNKDWDNDUúÕODúWÕUPDVÕ

%LU EDúND GH\LúOH NÕVD PHVDIHOHU DUDVÕQGDNL X]DNOÕ÷Õ KHVDSODUNHQRUWD\DoÕNDQKDWD GH÷HUL NoN X]XQ PHVDIHOHUL KHVDSODUNHQRUWD\DoÕNDQKDWDGH÷HULE\NWU,620$3X]XQ PHVDIHOHUDUDVÕQGDNLX]DNOÕNODUÕKHVDSODUNHQNÕVDYHJYHQLOLU

mesafeleri birbirinH HNOH\HUHN ELU X]DNOÕN EHOLUOHGL÷L LoLQ

KDWD\ÕNoNGH÷HUOHUGHWXWPD\ÕEDúDUPDNWDGÕU.

(OGH HGLOHQ EX VRQXoODU GR÷UXVDO ROPD\DQ J|PPH \|QWHPL ,620$3¶LQ JHQ YH SURWHLQ GL]ilerinin evrimsel

VÕQÕIODPDVÕ konusunda, evrimsel mesafelerin tahminindeki

KDWDODUÕbelirgin úHNLOGHGúUPHNDSDVLWHVLQHVDKLSROGX÷XQX

RUWD\D NR\PXúWXUAOJRULWPDQÕQ YH SDUDPHWUHOHULQen uygun

úHNLOGH VHoLPL ile \|QWHPLQ GL]LOHU DUDVÕ HYULPVHO LOLúNLOHUL

RUWD\D oÕNDUPD NRQXVXQGD YHULPOL RODUDN NXOODQÕOma potansiyeline sDKLSROGX÷XQXGúQOPHNWHGLU

4.

.D\QDNoD

[1] T. H. Jukes, C. R. Cantor, ³(YROXWLRQ RI SURWHLQ

PROHFXOHV´ Pp. 21-123 in H. N. Munro, ed. Mammalian protein metabolism. Academic Press, New York, 1969. [2] M. $ .LPXUD ³6imple method for estimating

evolutionary rate of base substitution through comparative VWXGLHVRIQXFOHRWLGHVHTXHQFHV´Journal of Molecular Evolution 16:111-120, 1980.

[3] 6 7DYDUp "Some Probabilistic and Statistical Problems in the Analysis of DNA Sequences". American Mathematical Society: Lectures on Mathematics in the Life Sciences 17: 57±86.

[4] I. Borg, P. Groenen, ³Modern Multidimensional Scaling: theory and applications´, Springer New York, 2005. [5] . 3HDUVRQ ³On Lines and Planes of Closest Fit to

Systems of Points in Space´. Philosophical Magazine 2

(6): 559±572, 1901.

[6] J. B. Tenenbaum, V. de Silva, J. C. Langford, ³A global geometric framework for nonlinear dimensionality reduction´. Science, 290, 2319±2323, 2000.

[7] J. W. Sammon, ³A nonlinear mapping for data structure analysis´. IEEE Trans. Computers 1969, 18, 401-409. [8] S. Roweis, L. Saul, ³Nonlinear dimensionality reduction

by LLE´. Science, 290, 2323-2326, 2000.

[9] D. K. Agrafiotis, H. Xu, ³A self-organizing principle for learning nonlinear manifolds´. Proc. Natl. Acad. Sci. U.S.A., 99, 15869- 15872, 2002.

[10]M. A. Farnum, H. Xu, D. K. Agrafiotis, ³Exploring the nonlinear geometry of protein homology´. Protein Science, 12:1604±1612, 2003.