疾患とDrugMatrixデータセットとの間の遺伝子発現の統合解析におけるテンソル分解を用いた教師なし学習による変数選択を用いた候補薬物の同定

全文

(1)Vol.2018-BIO-55 No.1 2018/9/18. ØCrg¶qZ€C IPSJ SIG Technical Report. ìñq DrugMatrix Ã”» ·¿ Ä q w w¨; Cqw wùrstSZ” Âï ¹ çür› ;Mh -£s` ¶6t‘ ” !:¬R› ;Mh©4 úw‰ >±y³Â1,a). “A• Ñ xSÚq Ì UTT” Óé ·µ pK“ z ˜cTp‹ f wÌ q …;› _nb ” hŠz ï³æ ¯ Ñ tGV s84U/d’ •o M” { ï ³ æ ¯ Ñ wËGÈvpK” Ï Õ”µ q æ ¨ï Å Õ”µ w Ñ txÕtq ytUK“ z ýhsMOq ` o Ùåz ¨; CqÓé Ñ • ç› ;M” Z€Uæ˜•” ‘ O tsl h{ ` T` z f w X x-£K“ ¶6pK“ z 7Œw úUŒ’ •o MsMìñ~ ª$•(tm Mo w ï ³ æ ¯ Ñ › æO wxÉ` M{ ŠZ€px¶ UTvo T’ h-£s` ¶6t‘ ” !:¬R› ;Mh ï ³ æ ¯ Ñ › Š` z 7Œ é Ñ • çT’ Ñ pV ” ‘ O tb ” \ q › è¦b{. Š` o M” Âï ¹ çür› ;M NUsMÔùp‹ ¨; CqÓ. Y-h. Taguchi1,a). ¨ÅQUÿMýF N=ùúwsgxÉ` Mz sr w=:. 1. xa Št Ñ tSMo Ì. ‹ K” { ï ³ æ ¯ Ñ › ;M” \ q pf w…;q. › yV` ‘ O q M O ¼ˆ Us^ •” ‘ O t sl o T. ’ q` M [2]{ » ï Í« wÏ. ï³æ ¯Ñ t ,nX Ñ. sign;SBDD) q =ùúwÏ. t xËGÈvUK “ f •x. (Structure Based Drug Det ,nX Ñ. (Ligand based. Drug Design;LBDD) p K” { SBDD x» ï Í« wq.Ï › itz =ùúq wM˜•” Å ¿ © ï ¬³ Û á è ”³ ã. \ w‘ O s SBDD,LBDD w=:› 4O hŠz ¨; qÓé Ñ •. çT’ w. sl h [3]{ ¨;. C. ï ³ æ ¯ Ñ U¼ˆ’ •” ‘ O t. CqÓ é Ñ •. çT’ w. ï³æ ¯Ñ. pxz ¤ =ùú›. FIT•ÞÃçˆútd)` hM. w¨; CqÓé Ñ •. çw¨ÅQ› Ug` z ÅhCqÓ. éÑ •. ç› Ëm‹ wx‰a ª$» ï Í« ~ ìñ› ž `. o M” pK– O q MO > witýF N©4=ùú› s. ï › æO \ q pÑ › æO hŠz MTs” Ôùp‹ Ñ ©. gb” { \ wMOxz ¨; CqÓé Ñ •. 4=ùú›. Ôb” \ q UDópK“ z ÿü =ùútv. O ýhs¯ µÄ UC\b” ‹ wwz f w…;xÑ tAb. ’ cz ¤ü •z ÖÓ½Å Ñ t&;DópK” sr wÕ. ” …;tz‚•yÇTpK“ z ‡ hz °S- ` o ` ‡ Q. t› Ëm{ °Mpz ª$» ï Í« wÏ U°ŒwÔùtx. yz 6S- b” žAwsM=ùúw›Ã”q ` o b;D. ‡ cÏ * › æ˜sX o xs’ sMhŠz f \ p' ^. ópK” hŠz f w…;0®LxôMq 84^ •” {. SU<Ul o ` ‡ O z ‡×t-‰ÛYUôMhŠz :¦ª. \ w‘ O sÕt› ‹ m¨; CqÓé Ñ •. xw=ùúå. ï³æ ¯Ñ. Ò å æ tf w‡ ‡ &;b” wxÉ` Mz s. r wytUK” { °Mzz LBDD x=ùúwØCiZ T’ ýF=ùú› sgb ” hŠz -‰ÛY‹ 0X z ‡ hz ' ^S‹ ôMq MO b:› ËmUz °Mpz 7Œw úUsMÔùz Á—pK” z 7Œw. N=ù. N=ùúq Ï $t. çw- q M. çÕ”µw. p K ” Uz LBDD q ‰a ‘ O t z 7Œw. ýF N=ùúUsMÔùtxz ýF N©4=ùú› ‰ p V sM q M O z LBDD q ‰a -£K “ ¶6t S Z ” v„T’ x€•‘ O UsM{ ŠØpx¨; CqÓé Ñ • çÕ”µw. ï ³ æ ¯ Ñ › z Tvo T’ ¶ U. •`. o M ” Â ï ¹ çür› ;M h-£s` ¶6t ‘ ” !:¬ 1. a). ¤ G¶g»¶æúg¶J fÑ, 112-8551, ÔŠ [email protected] ŠZ€xj¶æ q ` o Z[AˆpK” { [1]. ⓒ 2018 Information Processing Society of Japan. R [4–10] › ;Mo z 7Œw. N=ùús` p‹ æQ ” ‘ O. t¦Áb” \ q › è¦b{. 1.

(2) Vol.2018-BIO-55 No.1 2018/9/18. ØCrg¶qZ€C IPSJ SIG Technical Report. 2. Ã”» q. » ¿ § ” ürwALx» ¿ § ”ür› îæb ” é.$s. O. ž ç° æ ¶ Ü t ‘l o !˜l o ` ‡ O { \ \ p xôÍ›. 2.1 Ã”». Ÿ‹ür [11] q zy•” ž ç° æ ¶Ü › >;b ” { ôÍ. ŠZ€txz =ùú› d)` hMw¨; CqÓé Ñ •. ›Ÿ‹ürp ˜’ •h» ¿ § ”ürp x°`t ¯ ž Â ï. çq z 0Åìñw¨; CqÓéÑ •. çwË ÜžA. ¹ çx0¯Â ï ¹ çp xsM { Hl o z ›Ÿ‹Õ« Ä ç. pK” { ² tmMo x DrugMatrix(GEO ID:GSE59927). uℓ1 ∈ RN , uℓ2 ∈ RM , uℓ3 ∈ RK x7‘ sÊˆù˜dp ó. › ;Mh{ DrugMatrix xå ¿ Ä t. :s»ZK˜^ •hÍpèU. hq V w. w. w=ùú› d)`. +t S Z ” ¨;. CqÓ é Ñ •. çwz. d)™wÌ !=› âO$tGå` hGFÛsÃ”» Õ” µ pK” { ™. tmMo x GEO › Ug` z ìñ] q t&. ps‹ w› s` h { é.$t xŽ<wè“ {. £ ú. ñ• DrugMtrix T’ x=ùúd)Ìwå ¿ Ä wú qÓé Ñ • ñw¨;. ç (GSE59905) › z ìñ T’ x CqÓ é Ñ •. ’ •” {. ì wC. 2.3 æ»T’ wÂï ¹ ç^R ŠZ€pîMtÂï ¹ çür^ •” Âï ¹ çxz =ùú › d)` hMw¨;. £. ç› ¯qb ” Â ï. ¹ ç‡ h xÕ« Ä çq z 0Åìñw¨;. wú ì. ç (GSE57345) › ;Mh{. CqÓ é Ñ •. CqÓ é Ñ •. ç› ¯qb ” Â ï ¹ ç‡ h xÕ« Ä çT’ ^’ •hù RÂï ¹ çts” { ² D×T ×N. x DrugMatrix T’. ’ •” hŠ. ú$Ž‡™µ Ä è µ Ë• (PTSD)• DrugMatrix T’ x=. xj1 j2 i ∈ R. ùúd)Ìwå ¿ Ä wôwCqÓé Ñ •. › z j1 x=ùú› z j2 x=ùúd)™wÌ › ¯b{ ™. › z ìñ. T’ x PTSD Þ Ãçå ¿ Ä w‚. ¨; CqÓé Ñ •. 7w¨;. w. ç (GSE67684) › ;Mh{. Uœñ › ;Mh{. wCqÓ é Ñ • wvØ´ñ. £. CqÓé Ñ •. ×!• DrugMatrix w¨;. £. ç (GSE40435). T’ x=ùúd). CqÓé Ñ • ×!wñ. w. ç (GSE59923) w. w¨;. ç (GSE15654) › ;Mh{ Ä` X xj¶æ. ÜK” U [11]z ŠØpx» ¿ § ”ürq z owhŠ~ìwÂï ¹ ç› ;Mo. ÉU xijk ∈ RN ×M ×K p K ” ‘ O s~ìwÂ ï ¹ ç X › ßQ ” { \ wÌz » ¿ § ”ürx K M X N X X. G(ℓ1 , ℓ2 , ℓ3 )uℓ1 i uℓ2 j uℓ3 k. ℓ1 =1 ℓ2 =1 ℓ3 =1. R. , u ℓ2 j ∈ R. M ×M. G(ℓ1 , ℓ2 , ℓ3 , ℓ4 )uℓ1 j1 uℓ2 j2 uℓ3 j3 uiℓ4. ℓ1 =1 ℓ2 =1 ℓ3 =1 ℓ4. Âï ¹ çürwAL› –l o !:¬Rb” MO› †Ìb ” ($ 1){. swrswÔùxz ¨; CqÓé Ñ •. xz =ùúd)™wÌ ‘ Qq H× q ñ w. ç. w. ™. )› žt¬hb‘ O s‹ w› ¬•žAUK” { \ •Uìñ › 0Åq b” =ùúw‰ tmsU” { f \ pz =ùúd )™wÌ ‘ Q› Ëm›Ÿ‹Õ« Ä ç uℓ′2 2 q z ñ q H ×. t )UK” ›Ÿ‹Õ« Ä ç uℓ′3 j3 U_mTl hq ` ‘. , u ℓ3 k ∈ R. K×K. x›Ÿ‹æ»pK” {. ‚” { \ wq•UôM uℓ′4 i q uℓ′1 j1. U=ùúd)™wÌ ‘ Qq H× q ñ w. é.$tx uiℓ′4 q uℓ1 j1′ U¨¢ µ üÍ` o M” \ q › > `z. ËÐüÍ› ;Mo " " # # uℓ′4 i 2 uℓ′1 j1 2 Pi = Pχ 2 > , P j 1 = Pχ 2 > σℓ′4 σℓ′1. i t|=ùú j1 t )Q ’ •h P ‹z Pi. t| Pj1 › z Benjamin-Hochberg p ‹U 0.01 Ž<p K ” ¨;. ì‹p K “ z â Ç ” « stQ x. Íz σ , σ ℓ′4. ℓ′1. xªj. O4Y` o z 4Y P. q =ùú› ¬Rb ” { \ \ p. Pχ2 [> x] x¾:U x ŽÍp K ”. ⓒ 2018 Information Processing Society of Japan. ™). s” {. ›Ÿ‹æ»xÚ¦æ»p K ” { ‡ c z \ •xÌ’ Tt a O` sM { ` h Ul o. w. tGV X /)b” =ùúq ¨; wÊˆù˜d› sbdt. w7t ` o ¨;. [^ •” { G ∈ RN ×M ×K x ¯ ž Â ï ¹ çz uℓ1 i ∈. N ×N. D X T X M X N X. x ˜ j1 j2 j3 i =. ˆ0‹UGV Mqt. †Ìb” Uz ›ìŽÍwÔù•w¦Áx×ÌpK– O { A. p. ï ¹ çürb” \ q tb” {. ›Ÿ‹Õ« Ä ç uℓ1 j1 › _mZ” hŠtz G(ℓ1 , ℓ′2 , ℓ′3 , ℓ4 ) ›. ‡ cz Âï ¹ çürtmMo \‚” { Âï ¹ çürtx. xijk =. ‰±ï Ó çw ID p K” q b ” { \ •’ T. O { \ •’ q ì4$s¨; ›Ÿ‹Õ« Ä ç uiℓ4 q =ùú. 2.2 Âï ¹ çür. y•” ür› ;M” {. ›. ” ‹ wq ` ‘ O { \ \ p j3 xñ. 2.4 !:¬R w. [1] › €°^ •hM{. M– M– s. q H×. Ü›. owhŠ xj3 i ∈ RM ×N. ’ ›ìwÂ ï ¹ ç x ˜j1 j2 j3 i = xj1 j2 i xj3 i › ^l o \ •› Â. w. T’ x=ùúd)Ìwå ¿ Ä w. T’ xÄ` X x. CqÓé Ñ •. ç. w. ç (GSE30122) › ;Mh{. w¨;. Ìwå ¿ Ä w › z ìñ. CqÓ é Ñ •. ç (GSE59913) › z ìñ T’ x. w. q MO Õ« Ä ç. £ vØ´• DrugMatrix. (GSE59913) › z ìñ T’ x. CqÓé Ñ •. ç (GSE59894). 7w¨;. T’ x=ùúd)Ìwå ¿ Ä w ¨; CqÓé Ñ •. £ xQ. q MO ~ìwÂï ¹ çpK“ z i x¨;. x- ”µÌ - ”µiU\ \ px. T’ x=ùúd). CqÓé Ñ •. T’ x ALL ñ. Uœ• DrugMatrix. q Og.w. ç (GSE60304) › ;Mh{. æ ï ÍQ(B´¢ ALL£• DrugMatrix Ìwå ¿ Ä w › z ìñ. ç (GSE59895). ËÐüÍw§u¬pü. )pK” {. 2.

(3) Vol.2018-BIO-55 No.1 2018/9/18. ⓒ 2018 Information Processing Society of Japan.

(4) Vol.2018-BIO-55 No.1 2018/9/18. ØCrg¶qZ€C IPSJ SIG Technical Report heart failure r= −0.72,−0.82,0.51,−0.09. PTSD r= −0.75,−0.81,−0.30,0.50. ALL r= 0.94,−0.20,0.96,0.14. (D) vØ´tmMo x ℓ3 = 1 › z (E). 1.0 0.6 0.4. −0.10. 0.0 0.2. é ”Å ` o z KO pè¹› ƒ” ¨; q w°•SUôM¢ 4 Y P ‹p 0.01 Ž<£ w¨;. −0.2. › =ùúwÚ€wª$» ï. 0.2 0.1. 0.2 0.1 0.0. 0.0. 0.1. 0.1. 4th, 1.66e−09. −0.1. 0.0 −0.1. −0.2. −0.1. −0.1. Control Human Tubuli. Control Human Kidney. Diabetic Human Kidney. Control Human Glomeruli. Diabetic Human Kidney. Control Human Tubuli. 6th, 1.1e−07 0.2. 0.15. Control Human Kidney. Control Human Glomeruli. Control Human Tubuli. Control Human Kidney. Diabetic Human Kidney. Diabetic Human Kidney. Control Human Glomeruli. −0.3. −0.2. −0.2. ×!. 0.1. 0.10. 0.0. 0.00. −0.1. −0.05. Poor prognosis. Intermediate prognosis. −0.2. Good prognosis. Poor prognosis. Intermediate prognosis. −0.1. −0.10. −0.05 −0.15 cancer. ex. control. ex. 0.2. 3rd, 1.33e−03. 2nd, 5.1e−12. 35th, 1.7e−02. 0.05 0.00. 0.0 −0.1. −0.1. Control Human Tubuli. Control Human Glomeruli. 30. (F). 33th, 1.7e−02. 0.10. 0.1. 0.1 0.0. $3. min. −0.4 control. ex. min. control 0.2. › Enrichr tž ¿ Ó. −0.5. 0.0. −0.20 −0.25 −0.30. Non−failing. 0.0. é.$txz ²–…p¬R` h¨;. −0.2 normal. to time points 1/4, 1, 3, and 5 days, respectively.. cancer. −0.2. green crosses: crosses: ℓ2 = 4. j2 = 1, 2, 3, 4 correspond. €°b” \ q pz =ùúwª$» ï Í« › * b” ($ 1){. 0.2. 0.2 0.2 −0.1. Black open circles: ℓ2 = 1, red open triangles: ℓ2 = 2,. 25. Uœ 30th, 1.7e−02. 0.0. gular value vectors of time points, uℓ2 j2 , 1 ≤ j2 , ℓ2 ≤ 4.. ç›. 15th, 4.2e−02. 0.1. and 5 days after a treatment) and the first to forth sin-. ¨; › âO$t KO ` hÌw¨; CqÓé Ñ •. min. 0.1 −0.1 Non−failing. 15 20 days. 0.15. 13th, 1.7e−02. s correlation coefficients between time points (1/4, 1, 3,. Í« q *. 10. (E). ¹xqt ℓ2 = 1, 2, 3, 4 pK” {. Fig. 2 Time point singular value vectors. r represents PearsonŸ. V sM{ f \ pz \ \ px Enrichr [13] tî÷^ •o M” z. 2nd, 5.60e−03. 0.05. 5. 0.2. 0. z z~¯z ˜GÈz t|z hÌÀ. 0.1. ›Ÿ”Õ« Ä ç, uℓ2 j2 , 1 ≤ j2 , ℓ2 ≤ 4, w :{. Control Human Kidney. (d) 1/4, 1, 3, 5 Ô™). Good prognosis. 5. 0.0. 4. −0.2. 3 time. cancer. 2. normal. Ì. 1. cancer. 5. −0.2. wÐ ž ¹ ï ì. 4. normal. 3 time. normal. 2. cancer. 1. ›Ÿ‹Õ« Ä ç{ r x. q H°T’ H›Ì. 0.5. −0.5. 5. 0.0. 4. normal. Ì. 3 time. 0.5. 0.8. 0.00 −0.05. 0.3 0.2. 0.3 0.2 Non−failing. Ischemic. 1st, 2.69e−09. −0.5. $ 2. 2. 3rd, 1.14e−03. (D) vØ´. (C) ALL r=−0.75,−0.24,−0.58,0.78. −0.5. −0.5 1. 2nd, 3.41e−01. −0.15. cirrhosis r= −0.56,−0.78,0.52,0.36. 1st, 3.79e−01. 0.0. 5. 0.0. 4. −0.1. 3 time. 0.1. 0.2. 1.0. renal carcinoma r= −0.60,−0.84,0.54,0.21. 2. 0.1. 1. 0.0. 5. −0.1. 4. Ischemic. 3 time. (B) PTSD. 3rd, 1.01e−39. Idiopathic Dilated CMP. 0.5 0.0. 2. Idiopathic Dilated CMP. 0.5 0.0. diabetes r= −0.60,−0.85,0.53,0.20. 1. Idiopathic Dilated CMP. 5. 0.5. 4. 0.0. 3 time. 1.0. 2. 2nd, 1.65e−17. 0.3. −0.5. −0.5. −0.6 1. 1.0. (A) ú ìñ 1st, 1.21e−01. Ischemic. −0.2. 0.0. 0.0. 0.2. 0.5. UœtmMo x. ×!t mMo x ℓ3 = 2, 6. › f •g•¬R` h{. 0.5. 0.6. ℓ3 = 13, 15, 30, 33, 35 › z (F). ±ï Óç›Ÿ‹{Ä çw6{[$ ((A),(B),(D),(E),(F)) q Ì ‘ Q ((C)){ 6{[$wÍw P ‹x (A),(B), S‘ | (D) t mMo xz § Â °æ s<¢ ANOVA q sA£ w P ‹z (E) q (F) tmMo x4Y P ‹pK” { Ì ‘ Qw$ ((C)) t mMo xì. :› GL` h{. z z~¯z ˜GÈz t|z. hÌÀ ¹xqt ℓ3 = 1, 2, 3, 4 pK” {. Fig. 3 Boxplots [for (A), (B), (D), (E), and (F)] and time dependence (C) for sample singular value vectors. The. b” \ q tb” {. numbers above boxplots are P -values for (A), (B), and. 3. AL. (D) and adjusted P -vales for (E) and (F), computed by categorical regression (in other words, ANOVA). The. 3.1 Ì ›Ÿ‹Õ« Ä çw¬R $ 2 t -‰^ •hÌ. numbers above time dependence (C) are correlation co-. ›Ÿ‹Õ« Ä ç› Ôb { ALL Ž. Žx ℓ2 = 2 U7‹ Ì q uℓ2 j2 w. efficients. (C) Black open circles: ℓ3 = 1, red open tri-. :wˆ0‹U. angles: ℓ3 = 2, green crosses: ℓ3 = 3, and blue crosses:. GV Mwp \ •’ › >;` h{ ALL iZ x ℓ2 = 3 U7‹. ℓ3 = 4. j3 = 1, 2, 3, 4 correspond to time points 0, 8,. Ì. 15, and 33 days, respectively.. q u ℓ2 j 2 w. wì. wì. :wˆ0‹UGV Tl hwp \. •› >;` h{. 3.2 ±ï Óç›Ÿ‹Õ« Ä çw¬R $ 3 xz ±ï Ó ç›Ÿ”Õ« Ä çw§ Â ° æ —Ì QpK” { Mc•‹ nM o ¢ A) ú. 3.3 ¨; q =ùúw¬Rz t|ª$¨; w* q °A ‘. \ \ ‡ p wALt ,nV z ¬R^ •h ℓ2 , ℓ3 › ;M o z. ™sì UZo M” { \ •’ t,. ˆ0‹Uì0$GV M q M O ÚEp G(ℓ1 , ℓ2 , ℓ3 , ℓ4 ) › ¬. ìñt m M o x ℓ3 = 2, 3 › z (B)PTSD. t m M o x ℓ3 = 1 › z (C)ALL t m M o x ℓ3 = 4 › z ⓒ 2018 Information Processing Society of Japan. R` z \ wALt ,nM o ¨;. ›Ÿ‹Õ« Ä ç uℓ4 i q. =ùú›Ÿ‹Õ« Ä ç uℓ1 j1 › ¬R` z f wALt ,nM. 4.

(5) Vol.2018-BIO-55 No.1 2018/9/18. ⓒ 2018 Information Processing Society of Japan.

(6) Vol.2018-BIO-55 No.1 2018/9/18. ØCrg¶qZ€C IPSJ SIG Technical Report ¯ 1. =ùúw7Œª$» ï Í« q z ŠZ€w' ¬pU. (PF ) q. ËÐU. (P. χ2. ª$» ï Í« q w. wÑ Ÿ ¿ ³ ß ”wY¬. ) wAL{ æ• DINIES [12] t ‘ ” =ùú7Œª${. »• Enrichr [13] w “Single Gene Perturbations from GEO up” ‡ hx “Single Gene. Perturbations from GEO down” t‘ ” ª$» ï Í« * AL{ OR x¦ ¿ ¶z{ Table 1 Fisher’s exact test (PF ) and the uncorrected χ2 test (Pχ2 ) of known drug target proteins regarding the inference of the present study. Rows: known drug target proteins (DINIES [12]). Columns: Inferred drug target proteins using ‘Single Gene Perturbations from GEO up’ or ‘Single Gene Perturbations from GEO down’. OR: odds ratio Single Gene Perturbations from GEO up. heart failure PTSD ALL diabetes renal carcinoma cirrhosis. [6]. [7]. [8]. [9]. [10]. [11]. F. T. F. 521. 517. T. 13. 39. F. 500. 560. T. 6. 18. F. 979. 89. T. 10. 2. F. 889. 177. T. 15. 9. F. 847. 219. T. 14. 10. F. 572. 219. T. 8. 10. PF. P χ2. Single Gene Perturbations from GEO down. RO. 3.4 × 10−4. 3.9 × 10−4. 3.02. 3.8 × 10−2. 3.1 × 10−2. 2.67. 2.7 × 10−1. 3.0 × 10−1. 2.19. 1.2 × 10−2. 7.1 × 10−3. 3.00. 2.0 × 10−2. 1.2 × 10−2. 2.75. 1.1 × 10−2. 8.1 × 10−3. 2.91. vised Feature Extraction Can Identify the Universal Nature of Sequence-Nonspecific Off-Target Regulation of mRNA Mediated by MicroRNA Transfection, Cells, Vol. 7, No. 6, p. 54 (online), DOI: 10.3390/cells7060054 (2018). Taguchi, Y.-H.: Tensor decomposition-based and principal-component-analysis-based unsupervised feature extraction applied to the gene expression and methylation profiles in the brains of social insects with multiple castes, BMC Bioinformatics, Vol. 19, No. S4 (online), DOI: 10.1186/s12859-018-2068-7 (2018). Taguchi, Y.-H.: Tensor decomposition-based unsupervised feature extraction identifies candidate genes that induce post-traumatic stress disorder-mediated heart diseases, BMC Medical Genomics, Vol. 10, No. S4 (online), DOI: 10.1186/s12920-017-0302-1 (2017). Taguchi, Y.-H.: Identification of Candidate Drugs for Heart Failure Using Tensor Decomposition-Based Unsupervised Feature Extraction Applied to Integrated Analysis of Gene Expression Between Heart Failure and DrugMatrix Datasets, Intelligent Computing Theories and Application, Springer International Publishing, pp. 517– 528 (online), DOI: 10.1007/978-3-319-63312-1 45 (2017). Taguchi, Y.-H.: Tensor decomposition-based unsupervised feature extraction applied to matrix products for multi-view data processing, PLOS ONE, Vol. 12, No. 8, p. e0183933 (online), DOI: 10.1371/journal.pone.0183933 (2017). Taguchi, Y.-H.: One-class Differential Expression Analysis using Tensor Decomposition-based Unsupervised Feature Extraction Applied to Integrated Analysis of Multiple Omics Data from 26 Lung Adenocarcinoma Cell Lines, 2017 IEEE 17th International Conference on Bioinformatics and Bioengineering (BIBE), IEEE, (online), DOI: 10.1109/bibe.2017.00-66 (2017). t ‰™| ›yÒ • Ã”» ¶6 (;•¶6ÓéÑ £ ¿. ⓒ 2018 Information Processing Society of Japan. [12]. [13]. [14]. [15]. [16]. F. T. 628. 416. 19. 33. 532. 529. 5. 19. 1009. 57. 12. 0. 936. 130. 14. 10. 895. 169. 16. 8. 595. 169. 7. 8. PF. P χ2. RO. 1.3 × 10−3. 7.3 × 10−4. 2.61. 6.1 × 10−3. 4.5 × 10−3. 3.81. 1.0 × 100. -. -. 3.6 × 10−4. 2.0 × 10−5. 5.13. 4.3 × 10−2. 2.2 × 10−2. 2.64. 1.6 × 10−3. 1.1 × 10−3. 3.81. ³ ã Æç³ æ ”¶ )| èŠþ (2016). Yamanishi, Y., Kotera, M., Moriya, Y., Sawada, R., Kanehisa, M. and Goto, S.: DINIES: drug-target interaction network inference engine based on supervised analysis, Nucleic Acids Res., Vol. 42, No. Web Server issue, pp. 39–45 (2014). Kuleshov, M. V., Jones, M. R., Rouillard, A. D., Fernandez, N. F., Duan, Q., Wang, Z., Koplev, S., Jenkins, S. L., Jagodnik, K. M., Lachmann, A., McDermott, M. G., Monteiro, C. D., Gundersen, G. W. and Ma’ayan, A.: Enrichr: a comprehensive gene set enrichment analysis web server 2016 update, Nucleic Acids Res., Vol. 44, No. W1, pp. W90–97 (2016). Grosdidier, A., Zoete, V. and Michielin, O.: SwissDock, a protein-small molecule docking web service based on EADock DSS, Nucleic Acids Res., Vol. 39, No. Web Server issue, pp. W270–277 (2011). Ú yÊ| q-«™| I%þ°à| fi«1| •Š Ä| eyôí| ¤ ¤Â| Œ<Y2| "•èº| ªÍ°$| _ Êóè| bjÏY| q-> ˜| ~>óÏ| ³æN‰| C {“É| á,9~| !j2D| ¦Ú%| ›yDÉ• j CQ~KQ ×!t0b” bezafibrate •Ow ®Qq ð J:| | Vol. 46, No. 4, pp. 200–207¢ ¦ ï å ï £| DOI: 10.2957/kanzo.46.200 (2005). üú¬œ| C{Òí| ÿ>Ü%| yO|Ë• jCQ~ KQ ×!t 0b ” bezafibrate w Õ8d)®L| | Vol. 43, No. 2, pp. 115–120¢ ¦ ï å ï £| DOI: 10.2957/kanzo.43.115 (2002).. 6.

(7)