能動学習法を用いた創薬スクリーニング方法

全文

(1)2005−BIO−1（7） 2005／7／25. 社団法人情報処理学会研究報告 IPSJ SIG Technical Report. .

(2).

(3) *. **. ( *

(4) , ** . *. ) IT. *. "!. # $&%('*),+-/.10325476 *98:<;=25>?4@6BA?C1D:7E5!FGH476/I J@KHL,MN -9OP1Q<A ?*C@K"RTS/N(U&N%<A?*C7VHW,DYX!,Z@[3\]+D,:Ê_!F,G"476Bà (HTS) b- cd&e7f5ghYi1N M&6Bj kl +mn-po qr saV/t,Mu6v2HKwnxN -yA?C7V/z{e7f5|*}TS~V/Hv2HKw&5HTS K" *N(z{0*D,:@E_!FGnb no1+6v,2"VBSB"j5 a1%<*&KY-#$1+Y*,Q G ;*n* (G-protein coupled receptor:GPCR) VB.u02_4@6BA*?*C ^VtSB(j3[\ A*?*C1b 0.6%¡&¢Yi@6B£ 20 ¤(¥¦%<A*?*C@V9tM*§7¨<©uª!§«9<~ 1+*m-YU1% 20% %HA?*Cu%Y~ 1¬ - 90% %u[\ A*?*C@V9|*®*476v,2Bb*+¯"j¢ - a,+"°*±&K k,² SB 8 A?C (1 ³ M ´*µ K(¶MNu·H-¸H¹@KHº^6v,2/bo,+6v2"VBS/"j. !. . "# ! () !. '. ". *. ". &. ". % && ' +$. ' ". % && '. '. +. ,-. '. $. .. ". ""+. ""$. ". +0 ). "/. ''. ' '. " +. '. 1('. ' +'. " "" +. 0. '". '. '. 0. " '. 233333. " '. $. " '. ' ". ». '. "$. +0 !+. "/. " 0. ¼ ½¾"¿À #$&%E5!ÂÁ')Ã,ÄKnÅ*MN-Æ,ÇBÈÉ\9EËÊ(X1Ì^¨BD\HE5!Í2ÎWDYX!Z@[\ÏD:7E5!FG K,wnxN"ÐÑ%u[\ A*?*CÓÒÕÔ*ÖD,:@E_!F, G+ ²,×Ø i76BÙÚA?CÛbY Ü0nÝ Þ&ßK(àºu+ _á ¯6w@ËK Qx"j ²,×&Ø i1&[\ A*?*C1¬ ¹QHâãnä*åK,w 6/æçA%"R è29Q6E5!ÂÁ A? Céê,ë^SBN M^fH%mì"íî§YÈX,Ì¨Dï\]%Hðu+7_-*vi1¢+ ñ2S/N(ò*T2ÎóôK<õ&x NY~ö Ø i1N¯"j÷YøÍ·"¢*ù-1[\ A*?*C1¬ E_!ïÁúA?C7V/z{0,K ²1× 41û%<âãê,ë a1m9üuý −43−. ).

(5) á Ø 1 i N*mMQMYj*þ%1w@Q<âãê,ë,V/@û6YûKmn-"ÔÖD:7E5!FG %<Ã,Ä+*+¯6(ÿ3 f %<âãÙÚV<àº476 b6/jY4uQ h Y- 6

(6) ¢*+K<º i1"ÙÚA?C&%Hâã VBtTS/N-BÖ1%HÞ&ß K 9%1w@QHA?C&%"Ù Úí!1VHà º*4* i &%*QnMYâãê,ë*b+¯6Y¬ V YN- A*?*C@VH í |®4@6v2Bbn*Í29Q6B j mn-<âãê,ë K ,Q<âã Ù Ú 1) V9z{0K(àºu+¯6(D:7E5!FG/A?C*|*®a=2S/N-o qr sa V/t,Mu6v2"V/ !S (j "$# - D:7E5!FG/A*?*C*|®a,%"*uK G ;*n* Ò G-protein coupled receptor % GPCRÛ " VB.u02_4@6BA*?*C ^VtSB(j GPCR &',Ë_!, m (<$ $ ) % * l 0Q9.u01*98:<;,+7 2) + ·#$u%"R è2*SBNn*Í,2 i76Bj GPCR % -Y+u · .&KY-Ê Á/ªËÈTEH/-Á8T¨"B - /T\]F9 10 Q -M2 h 16 biogenic amine 4 V 35&%, E 6uTÁ925476HÔn%<umH2 $ 7*r $ " # Ø % 829 + :=Sf i,N¯"j -<â ã V9æ< · ; =K(àºu+¯69 Ô >&%HA?C2*SBN -1vBi % biogenic amine & K ? t4@6/ ? @ , E 6uTÁ Ò Ê A(F Dï\, ÊA(FD\ Û&VBt4@6v2HK S/"j 41Q h Y - Bó, E 6uÁ A?C2_ Ô CD*$7V9?u h Eu<Arnâãí!1 F!D¬ -"z{0Q^G E 6 × ÁúA*?*Cu % H V"-o qrs*aV/t,MN( !SB(j. I. ¼ JK ãu%(à ºT2,8¦ MÛâ $%HArnâãÅ4 w Në kOP b Qû i1NM&6Hí*!uF!D Pharmaprojects Ò 2002.03 Û 3) ¬ .10 98:<;1 % R S,+")ÍS- biogenic amine ,+6(ÊnTYX^Æ&E9(-9Ê Á/ªËÈTE9(-^Á8T¨""- D,¨<"V - U D WÍEH"-9 /\ÏF,n% X*&K ?t4@6 1461 A?C7V H × S 4) -4YZ Ò Ù Ú@75Û&2 S/(j Ô`, - [Z Ò ÙÚ&Q7S_Û mY Ô CD$í*!uF!D Available Chemicals Directory Ò ACD 2002.10 Û 5) ¬ Ö1% \]&+<A?C7V/|*®TS- 212914 A*?*C@< V [ZT2*S/N"ºu"j 8_2` 100 a 1000 K ^ ÿ b c_Ñ 6 def c_ ¥&m C, H, N, O, S, P, F, Cl, Br, I K V g ÿ b Ø ÐSBN hi i1NM&6A?*C@V 1 dK¢^2Hû6 4 ¸ jlkmn ¡ n A*?*C- omn ¡ n A*?*Cu m p7f qõÚ& % rMY2 s0 8 tì(!u4u7VBl ¡ vnA*?*C@ V p7f "$# -9âg á ã,b ¹&Q(2 P `,%<A*?*C &¬ - YZT2*S/<A?C7V9z{7wf_|*®1+¯6< aVB â w476 x 0+- ACD %<A*?*CÓÒ Å4 w N biogenic amine 2 ey&% GPCR *;Û V [2Z2S/NHtTS/"j*þ% +~ % T_- [2Z Ø û( - [Z&m( ~ z, K {7ÙÚ^4 V |b3S/N"ÙÚubQnM32 ü } i1<A?Cu+m*Qf2*S/NM<A?C& % -K,·BÙ Ú^< V ¶nA?CubH¡¢Yi,N M(j Û 0 ì(!u%Yàº ~ V8 - t215 ¥¦%< 8 tnìY!u Ò âã _*Û4 V s × SB(j MDL Molskey6) % 166 ¥*¦ C 7A2 r bY Ñ % 7 ¥¦ Ò ClogP7 , CMR7), Topological Polar Surface Area8), Molecular Weight9), ? nÑÛ Hydrogen-Bond Acceptors9), Hydrogen-Bond Donors9), OH 2 NH 2 NH2 %< L. −44−.

(7) Cerius2/Diversity10) +V&û(8tnìY!1% 42 ¥*¦ Ûo qr sa oqrs=29mn-YrsV Ò Æ n©"!1<ÛbYrs*í!u*VHo q,0K<|}4@6Hrsà+6Bj*í*!1V |*}4@6u2H¯K mn-æ^·/rsu%<z{& V f Ø E^612Î Þ u+¯6Hí*!u*V9 | ^w7K l Z / 2 G tVU Ø i,N M ² K,w(x 6/jr s1Ê" X Au E 2UYé,% £@4 V Ø QnM Qbag Ò Query by Bagging Û=2ËM a, b ,V - N Ø i,N Mu6 11)j !F, G+&[3\ A?C&mHÑ1 ¤ aY Ñ ¤YA*?*Cu%= 9 Ñ d Jµ&%<ð*?TS/ ¬ 2 ø E2 - Ô*%(o ÖqD,r :@sE_a& Ø V9çt4@6u2ËE5!ÂÁ A?Cu % ¡29Q6BA*?*C < · ¢¤ ç £ +61¦ 2 ¥ §3SBN^S ¢T]j,þv<+ -ÍG E ¨1nZEHGH47612<¯ K f %,w7 Q ¥ b % ©H V ª 4^w7_Q< ` «4 V S/"j ¬ Û |® Ø i1<A?C& % ® [2Z&K9¡&¢Yi@ 6BA*²Ø ?*± C*4,Nbn~ 0K<Ù ÚQ@S*¦2 ¥b Ø i11·Y% +mQMY4j [Z2S/NM<A*?*C % -n+ l - ¯°YZ S %<2 Ñ u& b ³*M@·Y%K(¶MV N {C*Ù Ú D@V/1M- ~ z1K<ÙÚ&b"2 ü }&+¯"$ i "z Ø {0,KHÙÚA?C7V/|*®1+¯7, 2 6v,2Bb*+¯6Bjþ&v"+ - [2Z A?C&%= ¯°YZ S ± %<Ñ µ ´ n u&b ³ M fj 500 du % -Y¬ ¶uQ Maybridge · 12) % 20 A?C7V/ | N*- Cerep · 13) K NÊ ÁBªÈÍE # 9 - U D W=E99 -B /T\]F9 1 % ¸¹?t@ V |2bÍSB(j/A?*C´*µum 10 º 6» + ¼ | bÍS - ½¾¿{,V % ÀÁu7 V &ûn(j. Â. V¼ ÃÄÆÅ²ÇÈ Û*r sÉ µ Ò ROC Ê<ËÛ Ì 1 K ROC Ê4Ë

(8) Ò Receiver Operating characteristic curveÛ_V/S/"j]í*!uV94,NHg7x^2H¯% ROC Ì 1 Km(í!1/Ñ&b 48000 ]&%72H¯ % ROC Ê4Ë m tV ÍUVüYZ3E9G"aÍ2_o qr sa,mY Ô Î476Bj Ê4Ë^VBSB(j Ì 1 +mn- oqr sa,m ROC Ê4Ëul b t ÍU¨1nZEH G"a@w&, fÏ@ K øS/NM&6Bj 4uQ h Y-<í!19Ñub ÐuQMnÃÄ,KYÅ*MN -HoqrsabuwuG ³M r s É µ7 V Ñ@SBN Mu6v2/bh*¬ 6/j L. TruePositiveRate. ÒÔÓÖÕØ×. ûüý. . ÙÛÚÝÜ°ÞÔßàÚÝáÖâØÚäã. þÿ. . . .

(9) . /. . (%). FalsePositiveRate. å 4 ¼ # & æ4ç èé mê¤Y1{ ÒëY Z2 ¥§ Ø i7[$Z,Ñ$ì U[2. Z ÑÛ -îí é m4ïlY{ ÒðYZ2µ¥l§ Ø i, Y Z Ñ* ì U2Y Z ÑÛ. å. ñóòóôÛõ÷öùø ú Ûô õ÷öùø /. 2 ¼. −45−. è é. . (%). m Uí!12-n%"|*} Ø i,ní*!u " é %"ð*?-í m<UYZ0-n%"|*} Ø i, YZ&%<ð*?j 43 " %<>! ".

(10) ÛWY! { Ì 2 KVWY! {1%H. Ì. ¬ tVÍUüYZ3E9G<auKwnxNVYZu% 90 ½%Yí*!u*VHà × 4,û Kmn-9U-,1%(í!u% 90% VHrsuK9tS9Q-.<i*Q QnM@v2/b*h¬16BjnÔ`,-Hoqrs a,+m- 90 ½%YZ7VHà × 4u%*Kn- U*1% 20 ½%(í!u*VBt4i$nM*Mv,2/bh*¬,6/jv"%uv29mYÙÚ*A*?*C@VQ6H3f0/3f ² ¶1.6 x 03 % 2&Kmn-Bo qr s&KwnxNnr sV/@x"ù b t ÍU¨1nZ EHGw& z{0+6v,2"VBSBN Mu6/j Û |® Ø i1<A?C&%<Ù2 Ú ® Ø ± [2Zu % -Y + ¯ØYZ S %< Ñ u& b ³M 20 A*?*Cu % { CÙÚ7 V |2bÍSB(4 > !1- 8 A?C1b biogenic amine V@G E 61Á/2_4@6 GPCR K"RTSBN 1 ³ M + 50 ½Vef, % ¸¹2? t7VBSB Ò l L Û j3viumn-o qr s*a K,wnx N fj^ K t : Ø i76BA*?*C@V9 | v2/+ - ³M(ü{,+"ÙÚA?C@V ²1× 4v2/b+¯6v2"V 54@6/4 > !&+6Bj 6 » ¼ 7 9;:=< 20 >-? @BADC4@EF ÃÄ 8 No. 1 4 5 12 15 17 19 20. +7V/SB"j. 2. ÎSÏ`ÐSÑ=Ò Ó Ô Õ1ÖQ×sØÙÚ Ó Û Õ1ÖQ×sØÙÚ Ü0ÝSÞ ÙÚß8àá â3Ú 86.9 75.0 73.2 67.3 66.1 65.5 64.3 60.7. 90 70 47 66 96 0 52 96. 42 0 3 0 21 0 0 79. 13 0 64 0 23 86 21 0. 38 26 55 0 60 20 52 13. G HIJKLNMOQPSRTQUVWYX %Z\[]^`_badcfeghQijkHlmNn 0 o 100(%) Hpkq Gsrst\uvwNxzys{}|Q_ vrPN~s biogenic amine JKfLHS }b|_vGfRc ¼ Å"¿. 10-6M. " ". #. #%$%&. #. Compound 1. ) '*. ' '. ( ). Compound 15. 1+*m-Hoqrs*aV9tM&6v,29K,wu-HP`,%HA?Cu¬ ÐÑ&%&[\ A*?*C@V9z{7wf_| + ®476Ha&VBü1ý+¯"j^viumn-9#$&%E5!ÂÁ') ÃÄ,KYÅ.6 Ð Ñ&%&[\ A*?*C@V<àº -nÖ K 9%,w7_Q<A*?*C ^V(D,:7E5!FGH4* i "Ý Þ&ß*+E_!ïÁúA*?*C@V ²,× E^6( ¬ 12ËM 4@K<R*4 " + Ø 6 a18 S % Í2_¶=2 Þ i76Bj -"0 8 tì(!1"3[\ÎVB¤ ¥ "|*}TS/8 N -7V/,M-<A?C2H ± , ² × 8@:<;=29 % ?t& K >476 ¯%6 0 8 tìY!1"[3\ V SË-<A*?*C|*®a1D % e,V94*4 û 6 b+6Bj ¡. um-£¢¥¤4¦$ , . 2_u¸ +"@xu·Y% +7-1¸HV@%¢=¤4¦$ , . %8§4¨ ©-3ª¬« ©-`¨ ©D- ®¢ ©°- Ø ¯¬¢ ©-YP± ]©-° ²£³ ©-°ó ©uK`µ²f¶ ·M7SB¢Y4 j1¢- 1%" 0ÊY*í,Ê` V ¸D¹ i14 º8» © , ÀÁ ,ÂÃP*rArn . -sÄ-Å . -¥¼S½ ¾@V¿M Æ 0Ê . V ¼Ç-¿¯n¢^SB «È © , © , ÊD4Ë¤t"D°¦$ . -Ïu¸9 . - n4É&Q¼S½-¾V¿&¯¢7S/ ] 7K"í!1 F!D%Ht7`V ÌD ¿M *5! * $" w N @ &ÍìY!§«9 0-pÅ4 § D ËU , . ¥ K ¶·M7S/¢Y4 j −46−.

(11) ã Çä-å 1)N.Abe and H.Mamitsuka, Query Learning Strategies Using Boosting and Bagging, Proceedings of the 15th International Conference on Machine Learning(ICML98), pp. 1-9, 1998. 2) Fauman, E. B., Hopkins, A. L., Groom, C. R., Methods Biochem. Anal., 2003, 44, 477-497.. æ http://www.pjbpubs.com/ç 4) GPCRDB æ http://www.gpcr.org/, Horn, F. and et al., Nucleic Acids Res., 1998, 26, 275-279. çèké0ê°ëì0í 5) MDL Information Systems, Inc., 14600 Catalina Street, San Leandro, CA 94577, USA. æ http://www.mdl.com/ ç 3) PJB Publications Ltd., 18/20 Hill Rise, Richmond, Surrey TW10 6UA, UK.. 6) Durant, J. L. and et al., J. Chem. Inf. Comput. Sci., 2002, 42, 1273-1280. 7) Daylight Chemical Information Systems, Inc., 27401 Los Altos, Suite 360, Mission Viejo, CA 92691, USA.. æ http://www.daylight.com/ç 8) Ertl, P., Rohde, B., Selzer, P., J. Med. Chem., 2000, 43, 3714-3717. 9) Lipinski, C. A. and et al., Adv. Drug Delivery Rev., 2001, 46, 3-26. 10)Accelrys Inc., 9685 Scranton Road, San Diego, CA 92121-3752, USA.. æ http://www.accelrys.com/ç. 11) Abe, N., Mamitsuka, H., Proceedings of the 15th International Conference on Machine Learning (ICML98), 1998, 1-9.. æ http://www.maybridge.com/ç 13) Cerep, 128, Rue Danton, BP 50601, 92506 Rueil-Malmaison, France æ http://www.cerep.fr/ ç 12) Maybridge plc, Trevillett, Tintagel, Cornwall PL34 OHW, UK.. −47−.

(12)