pvclust/0000755000175100001440000000000011551360416011772 5ustar hornikuserspvclust/data/0000755000175100001440000000000011551360416012703 5ustar hornikuserspvclust/data/lung.RData0000755000175100001440000025246611551333722014607 0ustar hornikusersBZh91AY&SYSpc|D1!$?x*{9x Q"*3P5P=O@fm4jt+T*Qc%,[Z( M nw吥U%7pE%JIE BN5Z@ (( )!TSAB+153gnJR!TmqҊQ*6[ 5uWj6V/ Yۀ((H(wTH$'-5{⊪}Jiǀ=(LLmXF4Ybՙ{(7}ْJ@)S)TX*JK *F{ ccJQ^m^ )h%HTIRU24h4ODDE@ A@#`?C!"$B@'zxCP?B"E)il4d=M m=6Pjd @P$Jhh=A6S h@2!@ 4M MM<ҍ6=!4y4Lڂ hd=@i#&MĄ {ݾQ-KRή o{kkfo/RW4X"K0ѭsUWk2r,z+] t*IvmDH͵ҵR$el`CȄ1K4ʍ®JȒ0Qi:N{lmPP%ֶ+IQp˥==]N5B^iWV^:=]&7⼊-xiTu3RBAQU2'liAfvk{F b&Pz՞Ӈy~G!!Ăv'Xc1s<{A٫ae5qg6]A1\d*2~ $HI) POL(*4e2&DD.) 3BF&E&"32*Q&$T$D&Ic;wi$,d0֧̒`f$)Q c`fns%$%&f1a0Q„HA$ن30S(ɀ&`  e ڭ"c3&"VK#H6 PJI&vti&`̈3` m6ddD#IaĠEtf)K!A3"XhؚSbQ!E!(S"H2 &,"C 3("A!1) RY dP@PBTيm !)#1E"HHI&f( DDPdML (L$I!$Y L b"4aĒ@1bvڬHR1L E%H%E0R&ebHfle&`da)(E$`1$0Z"1b&@"uu\)&aL]#E$eȲRIB``֯WdMmmU>xUdZTxT&A1jK-J7ߩP.*HŦXRS_ H֫ LMң%G^8QÏ$aAd70Iֳ2x w P"d䡗kHШUե t)\t{6v11))) 6YvejۏpQV.o臌}<+Mvaa]͎[}3ee7wG;ݴ]Z:7`gY?awI8G4iȔ֬٨ņcRH>m٢G;UbMѐnN 9"}[Vѣ-PU9<*jH"Ib9v-'˻$ld!ueZ__%/i6R1)S[Uw`Zə%cMqtk1&` o*5O7?wDZߝVw,\_dk`=&UJ, T}#Hi҃WXhӳІ%d)>(~?Aqh}]{E*`|V/xP[zTe|\~ySJkD5 pTUN;mH_X܊*7Sӡu-n6k"cE m y=ş s+yͷ]D!ܸ1 ;tBIw<&$vRj){xf;.(i;$@Ic ̅&5[+Ʌ5oYjI%eʪANZ+M~eR~T~8g| sϙ\jt<)}ٕ8$+]*6:6:i Z4"eNϓO3w}́Twˬ'N\aэtZ&(;߱=p;roݵds;LNϝ䌃פֿWfF!ŵ0 "A^r{P"3ZeUVqa Gʩ8yJ/,YéVs4O խ;$PdpKl5޻-^*V'~[GL=Rĺ/nb&l)KA=e&iwm<[횭i<4Ikdži ,Y wTM1bIAe}iX匕5vk5Mt>qA9F֝ƻ^;E ب(0]] 3f ^r|BdbKrB,w24j q ٣8z&A䔛]VK%Xo]= '[wH/thF@.̏baPH̺mbۊDaqoWZ(I9}L}y񇊴j TA$cy &?4uo JhEJ nD4`e9)H[Q6(; Q1 h*+j[r( SO݉ܮmO Dr? SnADϵU3QZkIqQ4!/h.u4">B6~3BPphzP;gDFraROIW$m%FȕBBtH"h="QTxˆX4t7ŧw.1JժյvRH(Q ,i ! 6֪?_یKejv?f*f$h")a*v] :PRe!:IZI@&>@SJWJvXԇ\o4FSPx%5/ Wy}5P2vBOE"}^4uD,l YV"+qIA m!BYnA_wӢVgu0QXRhvܰK ӔR&צRֵTG/Q ZUC5Xĺ 4(@qEHQ*U7e|x gV9_Rz~VX2˾/cg EAEnR:ޞ["G+, {.ERoi7J\LLrD ]Av8rđӕJO3<TIZ#s].--qkp.~!X0.Rh\j?Jw*(FZ?Un9B)"de҅s~b(<*w:L6i"FD{wM]=/;+JCʛ;}/ 'YeuF⍄L2(*ta'I4yЙkΪMb[̤# %#^y^5ҥ3\ZR֩<|;I9sYB5ו]M|U!m&ZWzRçvkv.KdrMF= )n4B[a,$"=H3G57(&&DN|D!ߚ|R d+YKQgEL/ދ?=Rn PGa7</w[C4N`Z2َ؄tK:]TUk2:?"-nMot*;N&Aӄک-iOۢU@˚ulOD[yk4"69&yGXZKgF9Z.+v߅ֶH B؁ yHI/cdljy"a:zhZǹ'w,J&V;0cH//Y2"^"bݽ]w9Xm~r2M#M䧉b (lzRʶ'X`Ѫj8 qkdo9v0{"RХQrS}Ċ&۟SbOsOLF חnk&uvcٵg%5*Eēwycي.f}/P>25 ųWd)k#(/ܦ4u -T&cdyLu,j亙{ss?ONٮ$nojTP wa;Ԟf*s"R{-@Rtٷ"mË :fATUnЁ5o3D`W&VI2qkz{9s\#ɀCnyŬR 8l#_D"W/1.dN7ZԷOuӛ'uR}(h1,|[KO4F`Ƙ'nw|]5iԎ5T՟[js+Xnx%/+ֺ,ne|`> Y**;!BUG컉Ǖ ]Fn.(8|"k579Q;fG9D/~f4@ȌrRU<¤0!P~W5]TqC+ضNx]9ͮxb!ůY6JP^G" \QA-h}:mPקx5 >{VBm$.MLҶ޳.sʗEY^TؿS55+g}v'9#KClP|b;! C 1+yU0c}nx'd U v|"ZzsHu i:#VfZFY X7BESy98L\NΞKHi("{8'roHN*Nb_\x:|WvcȂC(I7*9ō8%Nw<ج}u'Dl$́Np۲.o+-ŢTv!Bwuyo ;tmjI)P#6)\!4 Sf''Ng-,6jq&u:{Xo $3ԅ&1  lFl b~ߪ9;M*R:)sy/ߜ}:B!.׬!oJ O3σKSNH[qRs]Q6ܿ[;Ob-MUQm"C HSiVlrE:|A5rffys`5>, :,Kʼ0A7{WԽ">܋2ş mj*grmAEA%>ü ·bNJ2ha)-(TI!$Gnt%? Wܚ)b_ġoV!B(Q8.=4֤3R,jk*渚oSŊ0TLժoqTU Ńf*YeN٤:MdȂb(n|+jkiDBsd޴#AS3v\z7U&ຉ"P:f*f K+s*˙f؈{J`2eNޗH8A^T]`h! ^Iӫâu\wZ㩲5ҸEЊ1rBd3//@sV8&:8czOuc$VD[jɸ*Iժ͚zjN#D2vaNp"nn 3F[J$D!TzZ*'KB>LmtA5v!xޜݭ8pROb=YA *zzTKF7ם8 ڇiQ$>&^i|Y5Zq8gUC'47"_ۡ؛Q ހ[h$@4~ #E.cÖƏf&NpG[|;(zE32yrs&2ȃGlgt+_ [[?:$?#Lnn4*ex2ͲCYغR7sF¹_@51xRD~S8@ѥ`4A(VRaK²!iVM"&NiJ_ !.xzK<#Zh ?] 鑼ٚs;1UESV壄Y٪cI_G;6΅`o m] qVmi1Gi92GUvQT-ʄew5BY,Tw\X2gf |x+]OVIAU|Ŭw Mv5+;#NWEV$asBm1. FEzfYcׂU1p͙#9a_DJu۬ 663U*uL#T #5kEf*A5wDUsr/EC&wIP* ؁p3(Q#֜[{0O3X7PH pڇQVpI {ߵa3FO;2 jE:(NdQ ȵOIf-JjxX&~zrԏ5$/|C A8 1V\t/PnPvcH?hf0Fb0NG`f7e es{JNNbH(λgc"sOe4$?3 =wDneVj3v:'"d涥~'yxJ o ҿ&t9bWzB%((1Q|uф3>qJ %ngCjqaw :Jq ׍j׏4eBYUaV;|:}HII0WDҠшH=Z;Y"]WGn"Qфд9k HQѹZJ!V\EN9Qs99#C 9]t p/+"QՑhZ_J+7'j%ɋ'9haZѝ@KHJMЪ6Tzty*sG]$#jxds4AsoV[ s ֻIII]LL|2Mv2,y)+>Zې;j!i -q JD.en?X:a:RL!XX27o.jQEOzmR}\w;0e~ ;dL-!8O-ʟ&X- Lg>gQy_]A~ \B. MC8Qj.H&>`4rh/S{(gf! p J4 -[\0uP+Iʌ6Yfyvri?Ync=@J&]}/&w8'ͰJ:ʍ:PUKUXՏQuu95r=n"T=>}{ X%X$5Ԯޞ%õd)j:Lr_=Ԩ^.+2͝w1Py}T:xS֧o$hi'̐(n7Z>"hͯT˦ zq㦱 Juʐ+F|#Ъ%p5ZWriÊޛCw8Z1!_tK : i.W:o]pth𘎕4f1Vg4Z'Ӵ_(veCVke1CUdFHpv~(EGZyOwY2i$9TG8{m8`!u,^iRF!Y[sZNZI>ւ]+;~- 8gK:VGa^!@9CuqJBw.ɘK;["R.R۠Lj(/n̓ E\>AW!CKX+\yA@|WY{Xax@uڟ%-cs&5cp7,|t!rɢX)*AY?Bt}:*^,{UMض\{o~e-p; 6ȇwGk3-GeJiy('wuv3uw7)//#8[@Or4 W--҄uywrVXu,> =1seiSu! Ԕ=ڇ.w_cN O[!'ka}ɤ۲![7%,Ø>j!ݿtn·7cIB찿E*q%lʯ _ jry}$~Q*Wl齃# F\]ːtd鈿E#u6t)ƲDqy"br.-y#EdX.̡~{kjDo`/Zxp+LV N߲z yԯAkvnlWU*h#f^PQJ5yf0 R8: /ic wYa{H8:a~HL'U"M+Ia>X?j%Z%>jy +[+u7(7ZX VQ*%!NYȃ#:EUu+)޻>8LjPf{R2"y$(Zfqnu^2>=0d{jt?| K.]bm=^3d!sEˣYh'<w?݌d3J(U'[1$U +禍b?/)fZ*#d YɣݠS4ذ'jc%9nPyȪRqmHkǺj(pm_,:_w/ʥ~ й3@Bg-: V%7)D|+\PiRcq՝0W^NwsoQA:!Dqt]նrOgxp_:ShYQ;v XC[̷#uw]?ܙPYCNHUG\5m>iU31rxd (z73wQ6Lu)o)\A>xh3}z!x_+(:f4$ }>Q?O20v {w(pxXI9C*htv 8DC2>5S\Q;5Q6vq ؈.<sGxV^[?qS6ׄP۝*>fS%‹M'5%CAܠU4%=ʫ&% tysnX\1-?|8"U9r$ou'Yho8b!-iAY3(u mv|PO>iMwEٱP ऱ*Mыa.e=M5yJ>f=a#vO*h~nhLq;ӥ%jBó}$Cx}_z7HO]uM Xث{`Mm/j BO$7n|z8qD#wf#<%֫L+ϹsN[2,9,ѦJJt)H<䭖u ia3C{{u݉q[ ;R:z+? y>L9x "ƨHc3)(qr^57*ok>D9vw~3kD`y5wmBR Yf*sQ &^,xt26t: d.4q“~x;gH^ G{/u]2FsL>fgMB6bJm7ᶝ;ˁ=t?P됙o:>\8 a\K"H|@*<_(i1yz97KL8j#X(rG iɆ3FHV;j!zK]r\^t@v8?`~d_FH5 H .h.YKf~9bAZQ70d~|r~]%%TGb)_b!]WzmkIK8'?BkMOK+}mzo?ԌuÒ+vJ2՗iyH0Q;JBwcIB4o〮6;H0rRTDJ992nx>7?<lhYϹzZL|ŠQiA.r[u3q)1+LIePQ4UY>`%YviN׀/AT.pWt\Xsh!mۗ iӮeQ/ȹ&2"U2-]v}A)-S4Y2 BX$Ӕ+"N^b=k`2A|`aoۺ^?,0o>,;$HTx&*O[\ޥ7K@REEs]8NZ]` 1Q ?j1 |q[¯^Npb r9p}>$2$6PT߇ϷqBg{o~4$9uv<n6'϶HWڍrJ}{eT>)1Lڛya=Q}l,*ШcK?׎%`b|x:zݗk# qAr{Z,΅ܾN43>oX5 &ػF<`K݇D.n.,* .ΪGR}fDP_li|rH{ Dƃu仝XKӋ8 *\!zs@ѸWJ\z%X_[;;&:}wV4|] խ]IMNKt|ǰumE(Qpc`uz8u?$ Q鐭-30MxӏA#/-ǚL DHGo*BQXiVyfsT(0)NRvjEԭ~3u9Tˎm։~\+۸#O;>mB൓orHʵ+ڎἃ'> h_@ SGA> ƅ7v.xC- :G[_Z~3>hLܕC>{`m6=G4f t)@?NT C)x!Fم:I:"7LC;aU#4lxyG,5gyhMsM?>kkd .H59pgY`vC$Mu9i:hUC?֦ri>n/Z'?LpT㑎w[=yB]{T.w`J`H]G/*Kꕟԟp0qy}ئp6y5̦Ba0M4A~=hק`غ ι­+_-rq9p_UZVY_@jϏoMYʩq츔L/!Gl76'"gEEX٣+Y4Y~o/} a@b"A/;؅򑇂hԼњV^e7Ҫ:fh,NNn |tLEJStR}*Sù*Yz nP/1<zҁLtowo WQ)'wy5R ]+&f[11\eZ_ޛߴըW3n6ﰅWύ%wpX:V< S&^WzROY9fꮝG~IQ>Ştv?}iR`WnZ:boҖ/g{jdtߧf4 iiPuĄdXb[Ɵ8]-M bZRu'Mհ-^IO{C1zHm]1UMDegi5{Nti%LH5c̬»_Z\jVՅt5w8IJPL??Reu;fsm7~WmCc2 O|k&+D_kM3Rjw ƌ'2w)\q+]f187~n=:%&8Lˮر9c}ϯ4y[>}wN.FvL_)+iۤ 2+[rI&Թf/Bȷ7e#7L]{%FwK%S7<әG _= xqW+^M༖'s;,P/gf㑐PwՃ 'ؔDyo9ۙ% zt5{7spNmXUӐ <Lw7k6֕4dꦡ'|vv\ͅ=ΨUFi KEK-^1NN=ZIɺPegK'd_QrEaK20c7{5cl-[ſ: ,>p ŰHcPn}z(\,%w+]#W%Ba=~kʶ9Av5!(@ l Juڱ~rv5\2١whSo*Zv{䋴xD %8}W4W?n!ORq #EȖ!iSwQY|8Jx6ڊ,iN5>8Fc)\:2^ H>{S~ƴ OAb8\*])}yޏ<w`5Mw{osTB2_r, >B8?b8f}ΘR$G[jaF0Owc'Hl?Qөl~8S}mf8n#i"EQ7Vp+g {S:I:Zc9>ڮ'hp*kw+Rb:mm lv5LJဦY&D˴GK͘r}=L[w).XTN@Ɠm:N}%((R_蟼R߲tp }߱x7ɢ}>hyl CKRE)(j{EdA9:t~'IJ?zk#K^šgӄEg-8ՆëJMMm쾛vQZsB "зRٵ/pgЬ@+33AxyČUgk{tg?/rKI{S")z  {KDjsEυ[ۊڹ!m5K\*R)wPr.v"FЋugYۮ8oڣS2rdϯΆCT Q6mrrXX蜒–hDAG,dwu.W(} fwV#,1I Iۜ*|4?|2;8r&4uϹAe0coKҒRn'ޯe׋G.u;( /?[Cö{wNZRGavыq*^/a g:ۭA_Wze> 8#bp9d$"q;2/<&9ڕl's[Bzo`>әΞRK,+n:>8%J:IBrW\v~=lBGěme9sʎ 7p_X]mtyHw޹~|ܚƯW !-d# >oX2Egٻh=嚋8k6rA=^N˒{-._mQw YV)|Nc^ ۞0=q|89WZv5lv>gfƵ .# ~۶L:S[4k6'^T-&6lA ;BPr'Y4\@X 0zIAge}sIw90V߬}5LO,j?r{ա{d}4Ɵ43}6{ζ (ī# ot MpHbo==|'sӗ}QjJ.h T~j1b]2g1O07>+ljgo=I*&w?=Ig^e5B1Dw'G#SsRVsXAJӷ~:Uw9K"D],]AbV֦R[yl P"k+}7+OrZg ԵBtƠA'(?( KCOKC gV|K6o%^'eu4)@e?$.5ioB Udk_1i=.6$ѻ X,궖)_]7̭ok0i6'}b XҮk8,˻ \\_؋jqܝf7$R98l҇UnN<ð @NQiSٕN:vßBù)25#nƋgSϤiůrr/Ft`ס KtT¥3 NFmժ{nuzrɫ?Wm3gn<lCkVE491څagx]?Cm M.Grxuf.7co}]*S'$\* o5]{Πz}YŗzcPЀRL햆oHrm<n˸+k7{Kx/Ehsl5qsGXf:hKn*hL3(&GSBT Ÿޕغjd|w:?b0, ;}}/L|)i,l.73aɣjtg!lNFi٫&JۃA&П!<]ϕ7pPm!|o^$Py)Zk,&Gs|i+=BII\>}zH /`ȱ}j[oCPq$VIT>!c,ڎ÷>ݿ`&I~ZC].r%f'ۤ̚". M\Δswr {CZWYXvԜ"O'dS7N]`U ^u{M^> `_{;4# v:˻ﵝ؇}5>*MZT5nO<>;lnr<+VvX;!!ڥk:p]i pFzwSb=>ٓ#m{کn6^==(tsYfN:q))׸O!ua_mqъvOKgNP ~J ۉ.ǦH@Ξ:-;2x7_#K8{v#-qH𖒎0"^?NdGG_ҧTF3Jxbb/ܜn{ :ܝ }3cIOM {*p1'!Nץ53ƗG!uQl9i.CIJ~(N9UGIZuGy۵_YS1ץǖ4:ײ]H]Ůf8̅7kAj.؊kR|9y//;.Mg wb^q|h|"Qo%,6#DHmMDOW[#ޝ75DFWMKh#WC[a8SПq㜋qlDĭg&߬R^xmG9өmæ1M23sM1J8"ڧP9*㒱asܑby#%<5{@8=P򺝒ΎK;21ۑ`&:^8=>CyF܇%Cu'(IIFx]^7E7fNKfP.\NyzIQ5Wy[xZhZ<0ڋfXiQ3xmˉڃ̷{L0|0y&4Bp^4l }Ўų?Nt{؛RtfF}l7;iY 2}EMJ .^oΛX Xcs ?p꿔6+ y0׎* $)p:_6W7'{.[tfOu(/xUw]wyQdR̵9CiW6iAL!Grpԧr$ҕ+Aͱ<3}4.z%w8dJVGa'pe,g)XJqZ'zziO-fZV$(VwV ~6(}/-+񱹬cl{ x`]ݗkM/W 3%O Ja ntxg^_:Pp-=/$QҢN$?Ix#'<T X~ė<=12vu;˷萻Ȓ<:WgP(v̻ƫ]vvhe`+g5Ɓ1'(zQAַ4nyPE1|KsNWj0reܭEÝޫ>;/7^·=drmn+E ͆Uq^$+?N+|/mF1W㿷ؾ'ys0Z*.x!q2Ԙ]O_^uN3*Wm:;ZO}JpYw'˻&a`sâg7& ƮdT`> +GjTl篥*Xȫp?a5P*6/ Moh3?Js[k'Yxrj_9w &'L ,u Ubmu\1v4 I! U>Jknې0sތm.=qZE~;.>:~϶eN!u`S'} o(F13|ʾ]ѥ(/ ҺOc'Ru|J}E=Z6 .c39MѤ? Joǯ\IA2 [w<|Q&į,R<莀 `aa$+YU?v%M*QiNfrZmAڗ/lBzKP7a][X)iG]GZ*_[Gע1(-=D?cا}a;xTZ^CZ}yڿ"e &)]G'?E8־ k:ޥEjs\4_eh;kx!=qZ;1.hg={?{?gZsbݠn~xkp+\VSYRwWN{֒cJG,yk9{ɖ޽_jNh>nMαikmٕ}pm[)|cp#(<v2^({i*;5%!"mF9]I{Eqb]R2c 4yJCXxzb; 0&2D.XʷຍYT^3 q:g%!2UZ(7*<~ۘZrh-0uspuĜƓR/\~EeR}B-MGkڏ*OI;M`e+BT GRƆuf_>97O2{ڐoi%y1C6I!QCX ńwS7D7 Zc5Y{?:kn.ގG"Nu"qfNc&UVwllҫPwGel`n/GV&66+,5 7Cۧg٣u;< lέ7#zλR"h^1,Ղ*1@6I !{/0F$!Z3eqC>ÕT]GW+o*K2w86HoʖcMO%z~Zz@m:m @ ZCkӖJd3?/nYsا<ڶ𤖏ϯ8,;m B7k@tF'ؙipwK"佥IcVR7jY3iiۉ>2vhX"18͛³;pgj'FGEFp|SF+L\ob8e\֟jU`xs`P+aF(c:Y1ǻSn-ȉo{1l:ӚZ ; WoIb&]~Zi:#Ț+'GQrqf[ߐME|Ks~HW2ҠЌҁZIA)[x3@o6T?uy,qGpsY__opw툺'|#p_圶OrIG86 >'U8`今BE'; gnMڨf۪n_8G%Kni_Y{U7pTJ"M۱ܮU+EЖ/K?z7`C1D/ۼ=BםPh&B<[?> _*4x_c8V3tvtKk s2ȕ!S 9Bv6M]|p7DӴwOq+OR'yw8_eݔ9Ӈ~׃} vuƤq^v㓏>I-)Qwv+ѷk,N0p$\&aSv>ᘼϩAsy:h\\+Zƃf}"qe#+R%R _4bİ>ϥan [m(a!ZOo|QqM> ~h[ܡ\K ]p}p疂L>'nu=C((% 5XӬV%H/ލFMxgͻ;'Aխyү f$G ]]ϗt_owLY?[7(Cǘ.VvX=oZtg'9ZpoQXQsH' 0[<5/8+i} ?W:D'[UzؽFx:܄6-dgY bN][)MvnܳHy nUCe4whkҤg -粮-u\&czeŽ`v }8q\&jFºCbŦϗIs${g)J\Wـ]5GN.0^BR3L57>_} a}R8mޝo_`g|1OQ,aJZ+y~@ [:V[ߧޡ'^rژOaU2LNj/1h6_G{1Is)RأwI9N JƩq z!ЃPXr"ifOx0)8951X|V,xQo c+ỐX8V@MĴui|>gA a~s4rØ?PŸs~yDRq~t9W0U].1%J?gf[uM [x_:}Z {G{_RSimt}9EڴlSY;"io>>CިrAW`j>Lrdq5`i9cޠwYgT+C*+KooNe ~~t}F88ID agruB꜊ou cٷec sނ^=VqwVKɺr֘Pj+ڒ4p?m_V/^hĈ}\w-l:HE yY]jpc?vr7eU73P蛏ˉμWrr%־~AwQ+2cN1AbUOUDe{BFcϟ/PNncs鷽b%jcAVNzEgtqNan\ƎZ~>½7vXOpJp$5;LvEt}Bp sԷ22w(+?G.mVS'|n'zJ:H#l)jC:q얠q:Flۙ_XیG+P*/P)z)v7vps|פ[BJg9UfM=):x- n؆zHfOJ3Rf>,%[>i9}y`<֜8_.Z)yp:qm]r3քԜ\uj_=x\Z)^?&֚|EsU/D^9vz7D-ܼh ( p+~5<k3zGHnozs~GDP-n8Tr(r| Zd ,A7JEqQEAoO2K]'ӟze}-~WT6RvXԭ]%nSFvyL>~yW_&Vpk7E\\D,Y;Q^Egy{zBoǭk:w!:A?ʽh;adpzex\drGu1VwfWnQds Y!GཉF+#Zmw  ;l}Q^̼e4>}e3Nop*Fz@p#ǘI4{_Su i{9\Fn4 j5sq[}o6ݼzR!#g˽]ޥ6_ c$>M Az^̰hϚmPx>S ~/]"D.*u,rwuXlu ;We{H KO˱ꯣ8=@}w#YT檺tlоRcd""Fu.rKbRԔ'rm6]gO?MB2/mҋ+-%B'Ӳ W2H<8^K-/()SGXw?x.m{AaXWXru*qKsgb}'8$r;NYuaO;r=G2Utp"ພ rN?SqںO}'~>49cwYu5ҴٵL^T7œ'2)VѝҰD67/(ԧ7jǟv8y#u0{gT4sdV`g޼4qW9K©Z$0d.+%[` 8 D|}6MJ?0^&_ sjPuګ+UE3 N0He`~k>YlFFB'{[^*{i|"[ӹZ:‚ 8#_;.z5zy|yߵ;Uؒt?(Ny1~84w ~ļIli7ĠõÛϥ䈇&Foz\QKع̮UUUQ}>hu+?sLI?>U-.7(O%շu}NmߝOUכ ܜaOu>&䰫T87ex?ܜ)hpN8;j_`eפu[{B@HA~.y9';{u:$ִl]\Qe& Cu?oV2d9:VY}A=U٠oeه:ڌG '"hܙ★ n,|n*$D Gu^S.b e5: TE|'K>HW>\Т\P`KRU,jiHӲ j9IGm4uJXUJtѶnWU?SZ8mZSzET)QdTсղē'i֙,p^+(ҪHPMU=5PdP%BᓤL_X2 JHQ;C ÕPYa;\rOvcʞ!ţ4}ߦ/wxݐ*Vh|G{TA +WRڲhB@Fe-uLԻaaVARFQ'6JAP{s =1;ǒ+2}62X42F!A/fR^ȗ 8V,u=D-ř)YRT%us*M% Ldٍl{y6]+#(`%Wӳ$*{ r6 4;\AJOC*eRV(wsXv˴ sgrrCuउy"@LY]IT6z"я,9v}=3 "0p3JD1a ME̬A:so)$RMj y؄vjh06{XRwS1XOj)qw q btS0tōm\.4fje.bw+N|o`wk;:D[i5d\6u8+f4yi(:X.oCXp y2|zﾴzm6wbQιZ'eѬIvJh&"YM V.|3p/h؏ *bޘјFU:몭ay[zhZM4#JLJQ6*5+뻶b&֍bLU$IY6j6͆Rɶ([bƄUIPmWWխj_Y>H@/)Rs{/:P/KNQu_t&c;(?V,  Hwy')Vz̯$_;2Sz)m^<۫Oz6ϵ:>)(F4,ޗi8ԂuIrTe"6 Y4T&TSYGAQEnrbvEAL,0zGXnM[^@2ͅot>yt#J]qx3ɗfH+nÜ.b<ʴeK'L õLnřsn1~He`cj!)wt<#y+ve㴦kd )U*v۰?E!1 oE$]eS TqQ͂m<ɿ&.IW[kԂbQ97L3 ORT[:&Yn=PK2)|Y"<حx䬧 qe.˕,\7_ jc؆F:\- PSFiV(-XCNCh>.CuwVÐDWnD]STX4~8tTͺ#ၚy2?*`&+P$P=ϒIӑ9;oZ g*(~ _oؠ*;v6.pY_ ł03L:įH=7Ѕ/:"pIױ̉k$,̵pYYnUQ0nD6MNlaOs/ PCrilBȮeFЧ[.SǖhmWY}ǎRHs-Yd: A_ۢu<$d!3}@Q aNh- h@ n|*DbUvٞnCD`hu,04I>\ kc5xd+ ֲ"&0* 'tHcRYK0٬T#$E< fZΧJ-OfiBLݣiPֱًtQH!L^[h#d<$Ә- i[e3 lE܊ {6YU=[PҢ:d+H♾ W\m[CbefeNVBEnO/Aw_|Y,Ŵ\bf|"Yןe$`'J$GZ=_ćȨV-|/>t~D!-z揯P$[^O҆t'.,A9+aj"q9՗JNY hT)&ĐI2,uuch5Dq"cBfS8n tJn( MT jHEvg &$&\^~=V~u j|1zVDv.LőZI9?nO/}X~y_UGHM-W4g"|X2Zqkj",Ց]=p-gj<铣-_NeaLbBFKU3ݟZʔ$JBwnw2j]bLӽEz?PLl(~~X6|תU$!˩@Vǩ4]^n"qpaUiaa@p/9QslƏ-#HQEgsd"y? Xjr]{T}xg3}cϴ%cԻQH2YX>_)\q",PEY욧B1f1 -§7Hzea6uo,i J= +$_vn#4BʁY% Ꞙr$ i\-Q#1كمyT1ɼǀԄdu!\RQ xS,4w "7MTec@ɍ*Q_][?C<1GdķUYa jdR)ɚB࡞J}ԚJSYtA܅:0Pa&U"kkZ&͖S,J~EjmtM j:#tN,ڱxPVz0s⠲7+'X,|ks.`_kGGDFtlHqmV*jO_'h_*z=n+5Ó|Jq>shs'qCmR=/p$5}@!: þ8P~A+jWިZ7{ P`$$ JOqqI9N}.T%&£OVuePUoe?e8nV!y |A#* JR~%'9V$JַQ&17ÎdYA?޲Ri W!UPmLxBLt d2u1{CFMًr !w#C6. /3"שu±(;1؁c"d1zp.'v&e1]FnBUE<7ZJ0;Y/xOopK1g$ittAi7ɣшS@ѐ˽]0 DTe`Ě̇5ݷ06!D鿑UPB`EѸi6[*/[rx ȑBD]z\Q؋to~_?-H}(>$AάjlچE+ EDi1H6eD=`ib8Uy7VAmI{⸏]͸dw!Ɔ:X(Bk&܋.iuʮ6@o2/q9.r1~r0)'w}Oc#.u2 5,˳}F^ O/uָ֠I"3763u*QِB8fnխ@yiЌ#եGdBd)I*HFhs3X۴Yx"Cb**$֤;۫oɏoYOYSF&pD+cr :ҩaͻˑ0ɉ؅-$F4 OUץw*8hYEi"v0.kُ H\P9"4yoILaQeuQ,!uXzV28J[cݍnˡͶc`%"NG:f0/L /f Y*YmB8[N p=IHKHvzlbTwKu ݿfEl =3ȝ9~sfg|ۯ :O GrX|ǒ!9P џ Ν(MO\l\  at3k؄,{>Q `j 3[-9>LYqYWHȂDo[T*=+$23@ItYlESv_O?v!e^'IQ />$4S3쀯*F\kI7mCӄ4&dIf?9W5e:.N}$ppUsPe=j³et2ѱ%"tr2U(P1HfӛS̥zM s*.1V c-Gg漰;Kv'kY[Tꨮ|'Ռϴj}~!yb4)1&lpc9|vL!^&Q`< ѯUm5]̩,]wwy62ɮY! <2]vU6&f lYA4(wHw0= 8諎N+FCṄqfQ:,qj6=@ʅw*+TF4/S8|h3A<Jd֬RtU8=Y+δDI!].w}MB@lߗ7J?ؗA*C[ ;r V~^ +[+}_X5ZҭHSq X^s} dˡ-jC$l{Y_egbC]uj8QbA4JB! @LP{pʓվXA Z LڌǍX[)aQ! <(~gZr'gtzKRI¢)"NV9),.,Z{y#o^xL:Kv;QwLb)9oRMΧB8'&kwi8F7R.SSBb2s0^b-Rݾ^->IK'hP5a3+7>!D]xhMTZhOUW`Gz̲΋~׳G sTV G{֦n,PT Ծ{PQ@): 1n|9*"ǠZ(*<]]bPoǼr.g2lVgfꜼdX㜳빞TMH8`rϚ*I[S{/UmE/q#m=HBvoǁp@#;Q@5/DOe=RrHO* yƳ9[BƸVy9LT(]RZCE4TVUFe ^ѥ>Sy<{ۜfCF=L=[|UxE8VS Y0V?> ݢ[}f$9++J H=.M|<<&VDP\ۿWNCHy6UP}]7Tj]2Rn}.hf缵^Sΐ¤7zF['NZM2`$AS qLļzj;: ID;-K.憑Jh ?рeű4:#4u~#ADq Q5Vb :g<,bغ3iB3XJ3DG ,yfӾ$`+@ϒ<Aٜ&%I` [h$r'wJ}Nօ/H ߽ ozuc) 'Z4茺i6ɚū|gto_!<Ɓђ>)]*sa: wS;$'q&BWsTuvS 65l# j~ djw ߣuƠ#895 G( xByt RxfYםn ?fiX!r~+Lŋ-!. S LCҪ0A 7nvFnPUb|8IE ]5i7;™BQ`&gAHWwWPCX0kX)v(ﲣG|%r7bL>eTN1&Ӧu1vfNaЂMV\-e]T|\fW?*\R3 #elQM-Y`a(!}게8r1Ge.f{G]OU{> Zt/$u)i4r;6 ]څ5㱕R1S!+B&DsEXpP۷JenK:Y"SS0dj(\ q.֙m n2;q K ;;4OHX +FlK[W Ow :AMmnmDV*R7u 6xcUT0Y ;%Jؑ%MZ U2",=+™A)3ȠuapQf%tIB+lB꬞w:|7΅~>ep}T丠U/}ޏSEV ѼF*80eEy_ZPa P{^u:kpt^ p!Zf ղĈRe!%Sg 9dMQ S'@pD3^/EF|^rLT>5|8֟ޔ13|xݠa+>$+ʘUП;Dr]g,dg$:{dԸ݋p;l,*mXvs-6$?!Qt/?w qG>4Oy;͏v29݄9Rc=;³F UqwimoZKS:ZkeޫNm'FmWܯ"ς䱗xPpo\g%t=iG1ޕ^T}ܿ<+7BjqsB[?Pra{6*2_,?Q7e_I'uCw\h{ΰC`]Z@r#8l#Ӌ eHܡVO8j{%|ypK>GFVu\Z{φ*8_Y\Z/ٿIc|՞q5m\?ģץ7,%}{I+n\pj ^]s5 'W'wϲ{]so;q:V'߭ r #/1gJd^0 Y 4K ᷿أ)`JHOE|M]+q2dtK.fM2bӻ'&61i $WsG쏷j?Crd:SY9+ׂˀB4O`KGokUn/Z3;l)kӽjttB+5cشE5[u @=RavnT@o)AUq" .ejԤ0m"NkXL\VUqݰjzRKsu򖒻%c-=(L,3FTRSO,8iPHyc5_E,yY^u.,L#eu]0:ȱdlQ0خw]NQ:,=E9! (S;߼PlZ*! vIՇ#~7|"kLQB͠+E2]# (1O$jOl~yf7XJ(T~r׊:|!;IA[ouՓ0s>2޼`?eݎL{m:BKd=la 0+qYp.Ze#PjD4^@7JJC'nۡZd#g(h3eKZ0Y!!_v-v 0t/a4;3Y?JyIL=ut_ jCvYagRDr#$X2xiN)Pj~gpIA9YUڹPﰋwOyJG$'|֧MO88jO v T l^dgvL` Qo\r5sL٘E+sʱGήxiSɭĢU?/9}sS4d 5]B;Ea؄yyE7M}/ç:%fU7&FȂ`LX-)OI I]ڵQ^4 C!LHL)%HC2Hb0$)1d0)RI dPA)dRDIXFP& FHIvj{?k~OW(#AKSQ3ʲ 4!*9aE)ќ;J;^deUl<;+yiE">EpFl Е-`)w{ R9#WeoK/7mr>V7]\ߓ3`@r;y?K^T[;C266aF5u 驋h͠浰*FEcjoMVUf՜V1WI)Zyqm d!#k7zjy=berqrOcۂa}YÃ:p!f -GjEsQU|Ң6/Q1y2Bս%]HˍWbO'g_\NtNŲAƞl=Xh]g^ gYyk~o4+7QdRG$P81g[#{y"l;.R1gVڝmxG{nwiiZkҨQl-qkpcFȫSACۺ޳t=%Wܖ3|UFC (LR9ľfU<:ͰuLjpe9cگ ݵ}J A{}k$ǥ #H>X؇ߋ^5)=_/97dX-]Yڙ?SfD^T.m~Iz"mR0b󊨸 zrQ'&f:G)Q%C)Rwؓ`y/h_O1Iu1]V{Mww3 l|yf~ OoEneiu54e!Oy@wP&Ď #&xva u@FTj'145 !!Ÿ4qPt#z/u_`vUҞgVx412Jw/Ԡ(rAeZX!H%Nҍ9+=<+#۪{ j3E6eE}ew}!&FLCu 0\u9QtPmϢ3[)t3E`\0=Z1Q=RMU3oEuV!;K%/+qEըF4C p)gǗw<5adL`w,S9)Ԕ+ʮ>f.o$" wS2nFdM]B{$RESG>O. '%\y(&uHZ}\KuuYW[ky$C?֎sq(˨p[`͹T.F uzJZZ[q:Ik*'Lq'Pwuﰦr@lo2-#-1rUjp#C$c- Ƙ#5Sb%Gc {E ux?Ae*v!YJѫ49 bB3R:R Z_9vrybv!djWQۂHz"H%tR3YYt%0NDǭ hq(~? 1Yn_ $&]7_lע!Etؿ9=<.!*7Cy';?̦~;)83yZïuh: /Y%M /n"'uVA0%4-cv&Wvb:E~yP5MQ d%U/Fw¤8rUa ͖UO0ʴh9 a ]I6@aG&0jj;);(Lb_9#r]rCy34ehb̿3%YeXBHIܒ&HM 2"цlMc P60HRi E Mȅ@2&H( i!%$h2)"$k&HoG{@{G.$JoɧZҵIVv֕Ac-2$y7E4܎u[ ^ifOyʑ7I&nTo*szZ2 UNBT3G+[]yB߽{nDTBIlwB~QR?zbݬ-ߟ#;*i3`LDִ#X_v<<=XuEB9j`gfZi lb*^Eh՘r!BS짎n:;1;Ozs3`anV7]TdU(v+9 *< ]IO&{4kՒh+jObTMxƞPf)TTBD9Mʭ :}hɑBmIYaTrOh24I7 ]Iڐ~ʷzp6\9۵cb٩kOiR{1>&uɆݻB3>v5ѪZda2c̡5n 06@j1-Hj3yE0T4mrt&'cx܈Sdفˆ nv*x ֙7m\ 'k7T1!9}'|YNJ]Mc0A}c7 ^9iN4z-.l$:%h'BơD6 Qz%rjј iC_XvN҂РMݕnM|<:%Ҩ~ NqҐqP,H,(o[ÿO!x`B%176L,Y7W su뼶K{:4^5DW~*n &ܥ] 3_RnBV#)!klkc&MLT<e:- Y%!?s\ Q^j#``]A@u^bDH@l&FCE35$L)`4$A12HH% A4$ ((I(@&"Ql$Đa033$H$&2D I H$@`d1RbR( !A$RP YdI$1D``TؚA3F5dh^a1C)Sx{ǎsXM!+#DIhS\ aq_ NU #Ƕ!䉃;~P 5hE \k1I%*D:×_ҧVK4W튔!N|a53HA7`V0G,K_j&]jWqR-@4cnWݩ[;#[J.-fŽF rHM3+'WS۩Pr*b ]bb}kC!+`me5u2/ bf4{αQ7..$hˇWI /T! /ju'Uj*H!ʄjgbADpŁVcb5vh=M(%TW.DMCO$xwc39wysyhH]rwtjnPFMZOJsdp"NZWwJe0B/jBQFh l d{Pp!V09rs X\)vD;ew$&"dtRZ"Vrg1(i/N(H^xvj V 4>RkՅ{g Xم:Yq5* ЧwLZ׼U-t#n?{kZJsuoc$Lܜ1qdR8-G{_=jJ<|=~,jUwasRh#:\+8;.(+so1ვPrxbh̄v#N^;%I" S[t It:O5@c0Ϊ;xF:QuQ }&5EPdU(Q-+Eoce`Pm\AK)6Kxle:ָ MraX1wod>i9n϶+*F"hāvb,ysUbCVcnS|‚ujM3wKEh2*n&+k#W֣eV9vc˭Ln~5W"SK Rׇ\B+xo.koxPd4E'E\c]vbdׯ 7 c ʞG./Z> Gݪfd`_ZUP}5r)K[IUS1?2e9  d Q`9(BCNŏƸK(zv<ȼt#OsANZVUzb)ADTWa DWo<.^ w!ܭ% Xe`#w޵' +nRեv+ <wpS`fwΡ(G0A*.w J]9~14ҝ)jJ,rZh ֆ^5C6[^apw'Ƶ @lJu׋L>ǰf|& ڼWC'iMmZ@xY?-hW f:+Bsl;iCZĹV)Qgo9((U+VւCW =;;yl~QNY-]nœguM{L{p42YH5# aZQH_'dTu`UWt#%~nwUMQ9[D˭wDL ~BF)Ҷ1t 8"sM74ϵVٝd#dU˙8!WPTp=6+z.F\5ei 뵚8ӦEɘEajuVӨGpWLt=er-VVZ1E\" E3-;?zq*B8LYPŤVO}tdٯ&ɣ8.8vsi}yOhX\B/rpq!{^ s2_ib`u!׵|qyhZUSt. טeKZlˋ)7yYԋWԻc0~5bLCy.NC +P% TtC \Ǝ, h(Ko:TS5KA7,wzz׏dwi]Ic'_Eđށ,^/ԋoHސh;"->!Z;Z*S!I({lVۯeH¾ȨJζq|t~/ߵؗuo'.7K`G 1NȪO=®ar?27J6:Uˊ.Ad<ݹ!إ4~({<$$2h`wFzA/]ڰnWuʶӰ\ 7y# T)t4r({8v[҆='gPCq[?x9/h-'KԶ|_xg4$VkZiOX*{;u 6R*JJS\͢s8'9w8Gyu~_ 9|TQgYUi;=;R4Ñs?K> ;PrbkU{ZWs4j){9Y&bhJ) %ʉJF(ˊzHΨR;[Ӛ_Z9b{ob[hOht_>ߧ\La")bA*Zn{%8wo,*Y!% i-`_Gl9cOst"~9;]첂VaዪuoFԸWX(cw'WƚδuI|gX/+"_SkJ:L*Po5aI ɵGGO~>[h=5;kz1q qh lޫ-j#1zw+arGgGUǶApLrSBTJ_ltlL:a(m*Pwm[@2uo̥s掗]kTa(ӽnIVpHW.4c_3H1*5 /czfvu脜tVWj?һhݯW&2h*[D `aڽVb47"$7l*tڛVa;i6p]}&V&;TFfjyy>X,`YH{w|3ўaϭ0U|WRkcB &pT\r)m|<rPNꅽM/eX Ed K](k[ŚppEMy¾bd2G$z3i.]8bBv5GR-Nt[d6$9|Xߵ6&kepr{(]7wGE'?FYtUK꫞/&kS&zA^~%UkRu\ 4QP:^7 $1{;K,1Nb(Mݬ+ -jJɝrFVՂyHԻbA'Ҭmfgظ9Eͫ[YFzG3V"UF{nvR| y`}2XU>Ut5jcy~Zr&Iڮ|Q.VG|q.k_y5>Jn^F-볾qn@=b\|+p!b_L)$;Y6 J߉l_\m*yM~0t\R4KůDu_Z[v$A:v1&|{KJ7-W7zpאMZ[\ע6+RpJ8Y^"_S_3id#b}R۲^$kYh% OkB:o}`gPA5Q.8wnxywj ^ ! 驩s?nƠ 5%Kcuk-,ErwBN [Y"фMO?MGNop&AhxSݠѩ/e÷yM;0nwCa-XJƼvQiwa3|iXǽ٭4rQWeۛZ;LYr2;n3OὫw2PD^S,Hv,}P{lT#xs8;v AeƚW>Cgg'n}Mwz8?< ?xoZ[Ec8.b$ ,~DެAì{ixVxk(q(M3?ĵ~}c'yd-\Z ib"glau^~>!/=y^@(4vKL}\rWKn' !\< c=sU+-G{U! A/c}(fR= A Mԍg IZoX]>iBu=|u[:3QadH^{>^)O E *P ˊ UbZjHEI}Xg^gcnUY'B5)˥$g_;{߰%#j"Cő.xHfb- Bu?~r_ݝBi"0wN%WpST})jL7=(/UZqH;$Bt~&٠[ֽ/;LpY6^6UHq\]zvIOjPCE3}h;L\OHNrik`?hB>|(\ӀYYⶣ}?zw{үP f*[d}]Os{Vm-vogRwpo.G׊t4ܕ70J<fN0t|+I{މ/wЕ+\k_  a6ti~f g_ORqs[j"ciWxb>ue:N>OP>QCJ1 %PRBxtr*}ZFE ~& ݛ^Wp_kz~%tΛkb>6&_GVio==O œ*Yf# 6$/*Of2aj1m^O9#:(˯wXs0l|%zd]@Pt. oZVD"s9 O}F {^_Xr_ Y񃪏@&% >?K NN|KHpO5i_& d('g.v4;%RhRY/CSҤlџl)jQNt4*E]~9[z&}\R '}rאGַn"yi}ҟh]5%`mnútφ~ڼe!Oc0]o(ܦMlՙ',.֥.=GBrkN/zN4s'qD UTtZLV 9fk,~gJ91'1W'liBfp՛7 A\-O_[ktʭۈֽhJ|z"{MIf &[G:7w Dz~)KmҎe7MTr/kV}w+N0$>%4{tRlyI&X%L swVbHe5,!o.ԂDqǟM;JH=S+k^~ϱvyF;,'ۮkYkӴwW>*p(ėhO \CR;<\pni|u6`;a'vBaxa$kJrm=4׬}kV|L=# lon_'Arg_Хߖwc2 SkGPe#L=.粊G0lu]Dk&%5r5hM*S=;Utk8ZWw> $fyq[rhnF<<dž0:f4vhay׫Ji]Ѡ#n[cPby>-~|lӾr6)\ƾLk i#ù.O o[UǙj 8ZkfsTUKʛADKMg H}$SafscʰPWvث7ӣ 9q{ȷDLl|iY(kJn8ѧIzu%pHy w &%QsuKYд⤲1I?NyX+N6י "r3@Q0 n۴-oYαigpk!ƺ1VPqBm& jlXOX qg#z,s59G3(_/{Ȯ<|^ANμ:!/.9$9@ _0T,hQjQ)d/)__{|56GfÔ rp=8*f]R||GE.zM'8BaSgioYgF>O'Of5S@^J.Sukp]϶Grs )qLL;V/C|O}vlhIn[Z 5,_ێi~.}K%.۷۷f;}˛x:L9hvj&ݱnUYC"N1#!Sdc|$#(\;3 Ó $j ZJ5'>N\Rv,w{.kcډI)J[%0\[:M 9Ad2K W#~!iӝwdffO9cY67ܚ#J XwGAM8‹Vf㑝-]v JƳ邬[RЁ} :H3EI%Q`sW0aҲ-MSMNsr@esiKLCEgv#]?Ft):c1"{BiJSE !BMB&hȽx=EGJȜwޯ\}uYk9-H5̰P"24c&kV;6:[|d+!,!#,b kI.:sG卐/>^M.k+Ly3pooˉ ;K8M/{C /D!n!ᓯ. c anyt@[&qXVcƆ yM#2bgn 嶾أ:oTbXcaO,BHDˁ0Ԛ4*eI!fO%!Z[uE^pdyc`6A݇/tId7,[%lj~˸ٌg${Zj X)|7QŨ vApgS:iSiyh$#"DX Z!9m!L ]F6a9w~foO'8XC*]gt~R"bQP9ɺ?.ODZU`%|?!8%RzpZoY_4ށ6 tL4ȩuYik| ź<)NnZĮ8uYM8ID, JcOƝwj:+*V<ېZ ˽6 ^"eņㆠ>X&;j^EƘ<񎙽w־vu.}RX<[_OED`58+3=189{ u5N/Y^^lY~Sfԭ^T݆!RݞkR5qJy};%) cIj=qu.-5yb {t4-X HcxCC; oON^P8X2U}gX*lf/81<ϤHu^(Kf˹Ntff(I.AAƛ_O0U.)`tp7{KO@Di5\0u@]RoJN h8![}uin>U Jv k*(S6T~m1ͦJE9N^p?+} TfnQ|9EۭoiM7r͚~VY6u.;esZqRʖ*6\B$ lZ㻺ٽK))·{!Ar >*W pJJo%([nsj7OBa?Kan[113J}ЎT\jG):٥x,V_fi&h XGk?:>朚ZmgqWXn=vj4xMҟr ԥӂocwFF&U҃nΑ=vjCša#1\3!ܿA7JU9'nlG5Z/,ڣ.ҥ*vܮlGr~ FU)$r{fKq9}>\#&0,򑜠׺V>nҬU9 ZM}js7jBއ-9Aޤ[iڊ3 &Lz\$%xN)U"`fE:$ fʽVЅOLk(Q4Oq;ji'3 Zj;vmoM3͌.E =ƼWL7dw֡ýv0E2 HYZ7V|_ El|Ppo佇4{nFGW17/4K^77tG[T'#\U~/DGW bʷ$BO@ ßmRnlz>4hM]n'RF.ׇ6hNhO}3O]_6]߯$(X +M=Nz+*ύwMչ.ee`qdo"tW8ӪsNL1{tn"(X3S xӡY(t~ԟaaМ>lؔr< ҳBo>g0ֶ-dsFi(FKCЈ&>{pԘ6r6֗+ ~;qh~WYw.rݖeg8Z24 ]a~vE16y;lE57nڔ|VoF[ s{Շ] @'Xa٣!|4> aNwH7@u e,rUV`*מҖ$"ʃORUJP}]"c0OG:qr)hIEkעbFTt f׷J|Ţм O-OW'ЧOA8b/jt-9x/-CVu s/~zc|;To3FKשbZ(s>%K})0.`ϟ>>٨T2cgHGͳ:,nuI ueaurpB0KFݵ@d>mF^aa-k>LpS~O'Фh|A-sP+% 5 sH>:=Wt)`ӆDƷv?fqꚡI>f#N7/f;<`F w{7_`TH]cĶeH/;Ou~& PxMGQszp_ YWn #A=Y~xп:}GXy{p)~g#8U}#_X% ۤJlc}OMvO~H#C&w,4^qVhq {NBǝuOS8^4T]^Ë'#u潊uWXnBe N.im ?ܪRvڙ0gjwh[Kkuw&F7FƗmG8^O5 RS. c{PLM,6^ }HנL^Skʲjg ĭ}_ǝ{TZ҅LfglApD|(/~^~5(0t3F-ʨ +y}5kfw^EJ3_wq}ljj^Kύ${~WG7VpY>#$5l7]gnܔ^ `rн E FM" UvKGbX[ {f5 g'mk%O^PBƵi6Kc9DU;6CLE&NC%a]RGʜ:M r ]2~` z_* UI>eyb ~2\]Glv2덿p IqA>w>"Is8;WsOzЖ*V|kG_,ѹZTװ)+i+Xaw k f׽j4 hP+n4v 5uDDdN!a'@ү}Utd~|BGbtbwJb,ݜBNWG8rؽ1[aKpǎ8Oo3PӔڷoXimBZv{iq,ns{9c{ 䔚bRz4d=oI  4w0`RO|1]q%j7IIk?6HQZJ]t!xJơ9ou]z{w_ܿ}oeQoީjӏ0ۍr==r㚏q'g 5P]H/h^h*VRO'Mg#U bH0I?fv]¥ǁtJTIm o2n#`(_}>c]0sVݦXP2"(,#Xs9Bhe+ A^E P1?C{qW^-4%I[NFrNrw,Q| Pqy'TatW,%D^ 9zv}t~kP~=j['q_e \#ޯx0!?q|l-lzZ3}Jj'ߠ'9Iڔ;|D DF@֑B 꾒|E%~c4pص{BZ6NOvUlEsG>ƿG\qRZ+.cHⴼZ2윉a1{/8_q7}c |1hzշޤtݭa@)W#X o9 խ4;QVd:N7L1U6uS@D_#BAVm4MNo1}@b^@<S .i0Ww?PX5hlgȅL&fNx;B&ϡݗa(ltnm-qyW }N#ցFU+ɀ?Xojurj'|i{~[Y֐CunjMWo - л;nz{3t廓b1Hэ[lG:O(;S87dTu݀c޾ ܹ!IxZ3PG;7_+S (ZX-GG%"5ci<JYmޅwb1%ZTـ.2ZvlO>˩+Q'aiF&]= o?5y1G+N %~tCWg*bfzܒ J"SAp-d__a7𤟽}Yw"֞:NΨu]nZu9=艹2u=ʌzY%]m3h:i!v=y‰Vl.ZN#x Sb%)ܯF%P9Bj HEg$N{rW;*!ms~uT3NMԢ)5_f{l}{H&Y'i]uʿ6^[XvIɚwP]o%J>@ i鳻yY1D:#n?fI0^vP%;'a Ͱ=g뿮>7r2+1FtsG~荕D\џ⪯!`ۭ|Q'M7B 5C7ru{ëuTӫ>=.ڌ˂S\6Mkʳ|[x)Jgq=7;l)JL8uz=o㮕7Ӻ7 _^uQˮڈ]\}e. ϴ헅M;LBء~:5I;Jl°dG9H^ pXrN}/xm{8Y8{@q PlKV.[CK&2TNenLAﶃ})p=L PiL4K5riC;K= !WrfԼ_K񧏶+{ ]7DVv!3jn[]nEc1HIeʣL8#g6|]#)db.~#iymF!":e?hYٝ J. 1C{R.ǜkH 3| u ,0lpVy/H/ϖ%ir6Y=/qX8$XX}y5glgY# a=FB66]q1U+h絥rg|;+o;SthN_ƾ vb#*Gn=vنK=]A5|YT AWy"}o#AݱwNA޹xC9}HV"GBWvy dJp f `y{/#~rH`|R3Wߩ9ۖ[ý,M4yJџ#o$/)jLi9z! ѹ]ܸ"y8؉nt'As[?n9sjOlRl뗠 BO)%'1zrZ ta1@ O5^Pf f*xM8w~|2^Ʃ ѭ1mʼb_䤠0f(zhMYtƑrfgcvFP!9?1,#x;9_4w]x[ْw{Դ<$w$Xvvsw'A0746 Ɠ.}Kur\ڏ;x{փ^}ħhKr;mP".&0Vc>\YF;~Zz q+!%?)_ܵ_K* I;=a.[u/YZ~uES Ő|⹽?"'`Ltgk?yJڏ0 x0 ;{u pwc D'r)Z̟Ӡce_y_G8Bn;_3K._7ea!bs8t? ؾƝV-9 >//?rxw_kK^r wgX;e-Och7].8X 8ogOSŻ8qz+:PnMTp8#5-W3-n$;s?go^̪)" hT{mn-8f {6$Rt [\&#}r[:73}_' ([l M˶3EH f5)y8ܶ|%/"&Mҫ֧5l,rqZ\wJ]:#t_Izzl(}noڰkjc%մ5psxt*"z+OI@[ism2BsEZ=>uX!&Fk]*Hd50Nfb$.@7Ŭ'< \|_K& dTʜ=x! Mb$;lԚjk{Hb]TOJt"<FtyTR1ijtv`$_ZTT'n CiJcbY&U==+U@Ј[T*;ᭆ(<\®p!vMt~Lx>vTfKMݰhJ0,1,Q\QF `Y.уEzɹ1UU ~@⑞]ǣ9Ie5v/p)w>M!Oc9/_t9j4g y᧠i|&q H:!7F&C۽h{[\P3!Bm7KWuɬ(D]6^^Ts0jHgY39۪\0"ǬxX[0ɖG1ya/"L;mh,7<=g߆ M~3YD209 ǃc3 gRu!K(F.&/F5Z +ZBYL0Y $tӟZ թr1ѐn Q "ަe=l˱dR*)cVxOSJAxCV Hy)WN,N|iqn"gcB=&]C$d#]7C*d50VIVwՠGZ_K(p,b଺w4ؒ3&_syc\u4",!T {ݣHm10#\ox睺LǗ 9(\:N//< E)mvyWTD|ds&NJ#ӵ[3o8j&M{/kPCۻSe\0!|=r#zp?sՃedˊ,ݿyȇ VZwVqv⻺CM-$H7ӷ_*p[*8qVE{n!jy}ũȥgD mBl,sc+i'$ 鶚J02Bd$P,I&Q@`dQ"e!XȦ#!$PLQ3&J#$aa1I aI%Umaa"LFLP҉ D E1LK 14V Fib4ȅFe"I #&HHѤhFLC$c2J2$1I(DH"l̳K)HE$RaX$DɊJeM%44Ĉ)L$I!4)!Dƙ@`d&i%MFXI f ɉ$d" D` L1&!IEFD"FBf32ћ RE6`YD @ŌF$S h#6A$RL€$3#,ƔI#(X&2C  cI1I fM$JfIB%3 0j!aH1$13BdI2( fLX E)("0$D121I4H4) $SJOma)ɔEF% $E0X LabPLD 2BȢ#e2IlLH2%l1uVm{/۟lxY 8E?Pc8PIRL'5q%k ȴtT gJsCóhRhdBk[#[-u_9F!ƷvRwL y+&t]]ssW[P Tb&t &@E,N$H?lS,$"b(a]P&'xyZ6UD-i2Zk]ഃu:0xlF,tJ) X\<‡0%+ ;8,PAaW!cˮc9J>^_g,HEV j |6vBo";X BieUӣD 2i0I3KU^bpDgerg;BIp{F Q7B0/(IrHQLw `#ā8Qt(C|׶z]y?աy7FGZNš& (I4f?9µ4-z Y؇Wu a; htȂdq\Sn{۪%vY@)"Vn\z[9") ]@|P.hႩXQ9tQwZa9}3\UvDlֳ3$"i*ꁫy'.lH3Nʣ/'ru->㬫;@1v-)|sLt&iS+` e=SR)1(KkU-ku* [%3o Yt JRY%UIp\&LsʽXԺwvP[t[2߾ߜs *ȣij8bA:\ZX_wr II R-$jԵLPJIv7VƻPFժcFTeZMm]vD0RS<[8qӊ$$tɣ$mW$^dV%[?ZޖO`dzyZO"m(DK<|9Z+8[;XN+u51#B!eX]u5djUCK{S3wG"黫Dы0&k<sF:^k'Nw<5뻿Z# ` bЄL?4`#")L]qvG0qL)4x6DN:깗gGmM:9y"'L.MJӱCt @= ȉ63DHĈ(1Di1DM $$M$(1",2 c!e`0B"E 1 ːc `DĂEf D"MFIdb $b6Xɉ,&&f%&H("ɍ 6BL&M(h4H14d@2$0ȆIdd IȒR!2D3$BiC!a$(e 02(͊DDf$ifbfdd&-2$̐aJF)$#DBH" a4#L%&I!dPi ("Lİ  $jL1HE2S@! F1)I)D1HD2f"2HHd(AٔȊ&a B!Hłi (ƣdhJ 2a R#& dC416 (-6e!D 22fb4BCML!&̐B!$fDa T) c 40#ĉLQ, &$`M"H̲!L4THE0$PR"BH$LS&$0d( )ɠ "R&3b0"HF0) D%FDdЁQJa,bSaL &ɘJ" LfPB@ b1B$$HH1 (PْH2LIF   "abA%I4A%&!% Т3&aɉ RA( F,$dP), % !1 !œ0JZ1a Ff&$hhFIBQC&BS%M$#%"@ Y2%$Jl6BL(AL͘R@bD" T1I("%fld H3L$šE&E0SHF H&b!Q$Ča$bP43"l&61S02""`ML*I"HfbMa 4"Q"fb %"!H$ fc4!Li&J&R@LJl"a$AC2,3F̉ )h@ FI`HQ2/?k[jw^C>N`2?@('` Ӻҙ6g).g,;Dec67X[Tql, ]1ּ洪д \ 4$ܔhKEmQNҭ}uڼroCgDc ~R2z8Xr#jA/$)Yotːh]LO.gq2B1ݹ"Qj-g774fEzZTp |G-Y$o{2z5#~c*] ;jm7asՅGZ(%ٍDz{hi-V[&(qSɺ#&Gɽ*-JUrHN:$W4Hg ִ}m/b,Ȗ t,xY1$IC^ 3-N:Gˎ Leӣ%m !\dp g8.12A˷uL]+/P~dB>N#]wڐٛˤarx?s*h"-/3qITcٮQ`ܮE̳w@h)sG~hx^NgԍL~٭ər;VuhM&66cS1 &I0&H!&F(Cd&Rd 20TA&QbLy>LFBQbh3o-η =툻1فj][S\*C;iX7+bj{B$TF8^d$-M+\V͜;[Xy*e1G(f;k":ÀC4˻Zm;!lLB^8Vsd-6`AK/*TWU;Bى %5F_Vd8<,4]8lr"q*jǸ͘+sEZeJ( bPKf\3ZPE> ʟN 3͈K.\k頳m`jOwt S\N^ {Pڑpd_Lg}b3)J }U#˥>B+ΠZk̞ƞcB | p\=Α=‰ =3 RD(1|Xp&ɹ̣ ;bUvOU77_$D Xz6v>bڄ[3Q؆pnzw:hß[twxqf+-"=,M;9M0uع9e$!$FnDĤDĒ"LI DR6bARĆHBDDHXA!J@  BdфNڛMCbơY3c1D I34%1MI2AR"!4H̔iц3&#P!T P,Q(dYY3@(ARh0L0 4S@ #@f##e&1dd3)H$"A1QHJdD%&F"ifBI`)I0%(1BIS$ĩ # !0ȦI$0"D$a"2P$fBAM3 &,&AJ`d0Idjh2"d"fd@i, T2 H32FBI !"F %A1LFf"dR`h$I )HDS `Q)fb"LIS&i0`L QiD )DJh2e$4)0E,B!hQ3)BaJJ"dR$ @""F FI"Q"#Ia1QJ1IDa$$hJfYL$F)@4d҅2I$!̒i&F@(M$3FlJ,Da403H"JFP6$`1$Fhc3K$X !,JCB@Y0hJ) !IbS!"Q4E)feBb$ e 4!b Q#RA(P $LK! 0B$f )"`L16 JD$LHL M"LئRE1!,2i@%1 Ec0ę̲3 $f`lf`) d!2 3)&$ИI% 2h SdLJRbeB"D%1L 0H,R$I$R&LF!4Bѱ$1P4!,2 `" IfeFfPPR$aJR$"D)$j!e &l ($!2%i"E0M i0$(d̢D4JF1I(2f ,H H4!P!6)"Ld26! IDʘIJed F%)F3BdM!LL%d,ƙJ D`Y"@efeBM3c,؉L4$b2"RDa%f12$HZ(b(%6`" hDf&26jI DLhI1hHD# Rbh!LЉQhL$1` LȦ@b$"M,Ѧ@ ("ȚiAHDJ@fabY 0BLbIDf"iL "a#EiR!șb,3#B D#2$"bd$cAi0 ̉)2dD$D!4#&C2L44Œd&HaR $)F1QAa!4Ibi(( dIIA "44LFQ"" 2D%6$!$DR"@ɨDiQfcDB0a"!HeCE2L d#BI4E1ILi4JQL  !23 d6DhL&1Y!4PD Ͱ`hA* 2Ia(F) &Y2!#a(P&QUN "3CIhH%HEd"# ,DDFd DR #!)2$ D}~[$[9u{yJRW,Ҍsw}5 Tf.(uWy mXJ t#a$;MPw{S̭];)2BJ(3Q=4ka=OO ].j2i/B o,r*8Wst샺$i`!=OkK\״6UVavCnK~E򯣘L bQϐ-hա]dJ=*"?KkZ 5EW NT%igS yhj94mYLa|bSy~ )Q%X7wU3>G;Y -{#HFTT4&T~,Q?)ri>wAjƑ@^y,p%7y60VzJ&T}E *̉0ǧeＣZ#;2(mZԼa4:P|'YЦ_uP铹-qp2w`W6kΝNn }SzȐǥYgZhuz*#YDxw1y']#hd.-FllFL>5[xy80\K GRtKҿu^K$J"£ ݇i`AWڀ]Ϧ}S͙ +v_xiҘ\*] Oo8L|tNeEV^=M<O:\4Ii;3瑩mѹw)P?]xo|Hc$kjB[!L#غ8'o|Ü^4,Պ}T rFDW7ej̤vls'(r4LJbȊKEҒWZ(f<\avHHB]LI@3$ȅ4CP/,ŕ/H✊ŷvnGuRA{S3]e(ą[=fNidQoe\L!GnW٪ZoUw]]tCqb>4ʚZO Y/wffj& "xPÐ)3`w4;![pw^syY4\-.W5f.-f ֪M.Wz( h(i+XQ9x:eDXZޮN^o2rxbyٓ\ty D3#q5A}2 uG wۤ' hΩ':I NIIܩ\9"tpbM2ggRHɔ:C&wq+B-&̈́E֦)NƓTUaݓ=ޫWEZ,RD4Auv&,g-25ךTOCYQ-F [wjrB /e#Z6c'mQm]Q{j_Y/Wjau{&#|ؓ02d:1NU (FaE@wUf0PG% x@-sGp*+Iݑu$qD2znl$Se0MTu}yUxPo.y>`2ZzuLk 3'PX" )L>vP7(}b>LFƙYrٶ*:+? &~TȴySzۄߑP$UuSt-86AJR6"(~-YhSYPbhv Ez,RW\ihK0H )QS U-õ *V!]+Zc.@d-ɪ5 AF(۶ m8~Bj24^0C[akZr(ojʚt\!֥wyaדMW"g ^'gf̮9܆~v.0 #fqt$* &q=zzk.es"BuVLxg7qFd@wYĨgi UF@He^~_, Hidetoshi Shimodaira Maintainer: Ryota Suzuki Depends: R (>= 2.10.0) Suggests: MASS, snow, Rmpi Description: pvclust is a package for assessing the uncertainty in hierarchical cluster analysis. It provides AU (approximately unbiased) p-values as well as BP (boostrap probability) values computed via multiscale bootstrap resampling. License: GPL (>= 2) URL: http://www.is.titech.ac.jp/~shimo/prog/pvclust/ Packaged: 2011-04-13 17:44:46 UTC; suzuki Repository: CRAN Date/Publication: 2011-04-13 20:44:41 pvclust/man/0000755000175100001440000000000011151225630012537 5ustar hornikuserspvclust/man/lung.Rd0000755000175100001440000000277311147522220014007 0ustar hornikusers\name{lung} \docType{data} \alias{lung} \title{DNA Microarray Data of Lung Tumors} \description{ DNA Microarray data of 73 lung tissues including 67 lung tumors. There are 916 observations of genes for each lung tissue. } \usage{data(lung)} \format{data frame of size \eqn{916 \times 73}{(916, 73)}.} \details{ This dataset has been modified from original data. Each one observation of duplicate genes has been removed. See \code{source} section in this help for original data source. } \examples{ ## Reading the data data(lung) ## Multiscale Bootstrap Resampling lung.pv <- pvclust(lung, nboot=100) ## CAUTION: nboot=100 may be too small for actual use. ## We suggest nboot=1000 or larger. ## plot/print functions will be useful for diagnostics. ## Plot the result plot(lung.pv, cex=0.8, cex.pv=0.7) ask.bak <- par()$ask par(ask=TRUE) pvrect(lung.pv, alpha=0.9) msplot(lung.pv, edges=c(51,62,68,71)) par(ask=ask.bak) ## Print a cluster with high p-value lung.pp <- pvpick(lung.pv, alpha=0.9) lung.pp$clusters[[2]] ## Print its edge number lung.pp$edges[2] ## We recommend parallel computing for large dataset as this one \dontrun{ library(snow) cl <- makeCluster(10, type="MPI") lung.pv <- parPvclust(cl, lung, nboot=1000) } } \source{\url{http://genome-www.stanford.edu/lung_cancer/adeno/}} \references{ Garber, M. E. et al. (2001) "Diversity of gene expression in adenocarcinoma of the lung", \emph{Proceedings of the National Academy of Sciences}, 98, 13784-13789. } \keyword{datasets} pvclust/man/msfit.Rd0000755000175100001440000000560110207330420014150 0ustar hornikusers\name{msfit} \alias{msfit} \alias{plot.msfit} \alias{lines.msfit} \alias{summary.msfit} \title{Curve Fitting for Multiscale Bootstrap Resampling} \description{\code{msfit} performs curve fitting for multiscale bootstrap resampling. It generates an object of class \code{msfit}. Several generic methods are available. } \usage{ msfit(bp, r, nboot) \method{plot}{msfit}(x, curve=TRUE, main=NULL, sub=NULL, xlab=NULL, ylab=NULL, ...) \method{lines}{msfit}(x, col=2, lty=1, ...) \method{summary}{msfit}(object, digits=3, ...) } \arguments{ \item{bp}{numeric vector of bootstrap probability values.} \item{r}{numeric vector of relative sample size of bootstrap samples defined as \eqn{r=n'/n} for original sample size \eqn{n} and bootstrap sample size \eqn{n'}.} \item{nboot}{numeric value (vector) of the number of bootstrap replications.} \item{x}{object of class \code{msfit}.} \item{curve}{logical. If \code{TRUE}, the fitted curve is drawn.} \item{main, sub, xlab, ylab, col, lty}{generic graphic parameters.} \item{object}{object of class \code{msfit}.} \item{digits}{integer indicating the precision to be used in rounding.} \item{...}{other parameters to be used in the functions.} } \details{ function \code{msfit} performs the curve fitting for multiscale bootstrap resampling. In package \code{pvclust} this function is only called from the function \code{pvclust} (or \code{parPvclust}), and may never be called from users. However one can access a list of \code{msfit} objects by \code{x$msfit}, where \code{x} is an object of class \code{pvclust}. } \value{\code{msfit} returns an object of class \code{msfit}. It contains the following objects: \item{p}{numeric vector of \eqn{p}-values. \code{au} is AU (Approximately Unbiased) \eqn{p}-value computed by multiscale bootstrap resampling, which is more accurate than BP value (explained below) as unbiased \eqn{p}-value. \code{bp} is BP (Bootstrap Probability) value, which is simple but tends to be unbiased when the absolute value of \code{c} (a value in \code{coef} vector, explained below) is large.} \item{se}{numeric vector of estimated standard errors of \eqn{p}-values.} \item{coef}{numeric vector related to geometric aspects of hypotheses. \code{v} is signed distance and \code{c} is curvature of the boundary.} \item{df}{numeric value of the degree of freedom in curve fitting.} \item{rss}{residual sum of squares.} \item{pchi}{\eqn{p}-value of chi-square test based on asymptotic theory.} } \references{ Shimodaira, H. (2004) "Approximately unbiased tests of regions using multistep-multiscale bootstrap resampling", \emph{Annals of Statistics}, 32, 2616-2641. Shimodaira, H. (2002) "An approximately unbiased test of phylogenetic tree selection", \emph{Systematic Biology}, 51, 492-508. } \author{Ryota Suzuki \email{ryota.suzuki@is.titech.ac.jp}} \keyword{htest}pvclust/man/msplot.Rd0000755000175100001440000000075110206613616014357 0ustar hornikusers\name{msplot} \alias{msplot} \title{Drawing the Results of Curve Fitting for Pvclust Object} \description{draws the results of curve fitting for pvclust object. } \usage{ msplot(x, edges=NULL, ...) } \arguments{ \item{x}{object of class \code{pvclust}.} \item{edges}{numeric vector of edge numbers to be plotted.} \item{...}{other parameters to be used in the function.} } \seealso{\code{\link{plot.msfit}}} \author{Ryota Suzuki \email{ryota.suzuki@is.titech.ac.jp}} \keyword{hplot} pvclust/man/plot.pvclust.Rd0000755000175100001440000000544211151225041015507 0ustar hornikusers\name{plot.pvclust} \alias{plot.pvclust} \alias{text.pvclust} \title{Draws Dendrogram with P-values for Pvclust Object} \description{ plot dendrogram for a \code{pvclust} object and add \eqn{p}-values for clusters. } \usage{ \method{plot}{pvclust}(x, print.pv=TRUE, print.num=TRUE, float=0.01, col.pv=c(2,3,8), cex.pv=0.8, font.pv=NULL, col=NULL, cex=NULL, font=NULL, lty=NULL, lwd=NULL, main=NULL, sub=NULL, xlab=NULL, ...) \method{text}{pvclust}(x, col=c(2,3,8), print.num=TRUE, float=0.01, cex=NULL, font=NULL, ...) } \arguments{ \item{x}{object of class \code{pvclust}, which is generated by function \code{pvclust}. See \code{\link{pvclust}} for details.} \item{print.pv}{logical flag to specify whether print \eqn{p}-values above the edges (clusters).} \item{print.num}{logical flag to specify whether print edge numbers below clusters.} \item{float}{numeric value to adjust the height of \eqn{p}-values from edges.} \item{col.pv}{numeric vector of length three to specify the colors for \eqn{p}-values and edge numbers. From the beginning each value corresponds to the color of AU values, BP values and edge numbers, respectively.} \item{cex.pv}{numeric value which specifies the size of characters for \eqn{p}-values and edge numbers. See \code{cex} argument for \code{\link[graphics]{par}}.} \item{font.pv}{numeric value which specifies the font of characters for \eqn{p}-values and edge numbers. See \code{font} argument for \code{\link[graphics]{par}}.} \item{col, cex, font}{in \code{text} function, they correspond to \code{col.pv}, \code{cex.pv} and \code{font.pv} in \code{plot} function, respectively. In \code{plot} function they are used as generic graphic parameters.} \item{lty, lwd, main, sub, xlab, ...}{ generic graphic parameters. See \code{\link[graphics]{par}} for details.} } \details{ This function plots a dendrogram with \eqn{p}-values for given object of class \code{pvclust}. AU \eqn{p}-value (printed in red color in default) is the abbreviation of "approximately unbiased" \eqn{p}-value, which is calculated by multiscale bootstrap resampling. BP value (printed in green color in default) is "bootstrap probability" value, which is less accurate than AU value as \eqn{p}-value. One can consider that clusters (edges) with high AU values (e.g. 95\%) are strongly supported by data. } \seealso{\code{\link{text.pvclust}}} \references{ Shimodaira, H. (2004) "Approximately unbiased tests of regions using multistep-multiscale bootstrap resampling", \emph{Annals of Statistics}, 32, 2616-2641. Shimodaira, H. (2002) "An approximately unbiased test of phylogenetic tree selection", \emph{Systematic Biology}, 51, 492-508. } \author{Ryota Suzuki \email{ryota.suzuki@is.titech.ac.jp}} \keyword{hplot} pvclust/man/print.pvclust.Rd0000755000175100001440000000306510207117642015674 0ustar hornikusers\name{print.pvclust} \alias{print.pvclust} \title{Print Function for Pvclust Object} \description{print clustering method and distance measure used in hierarchical clustering, \eqn{p}-values and related statistics for a \code{pvclust} object. } \usage{ \method{print}{pvclust}(x, which=NULL, digits=3, ...) } \arguments{ \item{x}{object of class \code{pvclust}.} \item{which}{numeric vector which specifies the numbers of edges (clusters) of which the values are printed. If \code{NULL} is given, it prints the values of all edges. The default is \code{NULL}.} \item{digits}{integer indicating the precision to be used in rounding.} \item{...}{other parameters used in the function.} } \value{this function prints \eqn{p}-values and some related statistics. \item{au}{AU (Approximately Unbiased) \eqn{p}-value, which is more accurate than BP value as unbiased \eqn{p}-value. It is computed by multiscale bootstrap resampling.} \item{bp}{BP (Bootstrap Probability) value, which is a simple statistic computed by bootstrap resampling. This value tends to be biased as \eqn{p}-value when the absolute value of \code{c} (explained below) is large.} \item{se.au, se.bp}{estimated standard errors for \code{au} and \code{bp}, respectively.} \item{v, c}{values related to geometric aspects of hypotheses. \code{v} is signed distance and \code{c} is curvature of the boundary.} \item{pchi}{\eqn{p}-values of chi-square test based on asymptotic theory.} } \author{Ryota Suzuki \email{ryota.suzuki@is.titech.ac.jp}} \keyword{print}pvclust/man/pvclust.Rd0000755000175100001440000001675310454440040014544 0ustar hornikusers\name{pvclust} \alias{pvclust} \alias{parPvclust} \title{Calculating P-values for Hierchical Clustering} \description{ calculates \eqn{p}-values for hierarchical clustering via multiscale bootstrap resampling. Hierarchical clustering is done for given data and \eqn{p}-values are computed for each of the clusters. } \usage{ pvclust(data, method.hclust="average", method.dist="correlation", use.cor="pairwise.complete.obs", nboot=1000, r=seq(.5,1.4,by=.1), store=FALSE, weight=FALSE) parPvclust(cl, data, method.hclust="average", method.dist="correlation", use.cor="pairwise.complete.obs", nboot=1000, r=seq(.5,1.4,by=.1), store=FALSE, weight=FALSE, init.rand=TRUE, seed=NULL) } \arguments{ \item{data}{numeric data matrix or data frame.} \item{method.hclust}{ the agglomerative method used in hierarchical clustering. This should be (an abbreviation of) one of \code{"average"}, \code{"ward"}, \code{"single"}, \code{"complete"}, \code{"mcquitty"}, \code{"median"} or \code{"centroid"}. The default is \code{"average"}. See \code{method} argument in \code{\link[stats]{hclust}}. } \item{method.dist}{the distance measure to be used. This should be (an abbreviation of) one of \code{"correlation"}, \code{"uncentered"}, \code{"abscor"} or those which are allowed for \code{method} argument in \code{dist} function. The default is \code{"correlation"}. See \emph{details} section in this help and \code{method} argument in \code{\link[stats]{dist}}. } \item{use.cor}{character string which specifies the method for computing correlation with data including missing values. This should be (an abbreviation of) one of \code{"all.obs"}, \code{"complete.obs"} or \code{"pairwise.complete.obs"}. See the \code{use} argument in \code{\link[stats]{cor}} function. } \item{nboot}{the number of bootstrap replications. The default is \code{1000}.} \item{r}{numeric vector which specifies the relative sample sizes of bootstrap replications. For original sample size \eqn{n} and bootstrap sample size \eqn{n'}, this is defined as \eqn{r=n'/n}.} \item{store}{locical. If \code{store=TRUE}, all bootstrap replications are stored in the output object. The default is \code{FALSE}.} \item{cl}{\pkg{snow} cluster object which may be generated by function \code{makeCluster}. See \code{\link[snow]{snow-startstop}} in \pkg{snow} package.} \item{weight}{logical. If \code{weight=TRUE}, resampling is made by weight vector instead of index vector. Useful for large \code{r} value (\code{r>10}). Currently, available only for distance \code{"correlation"} and \code{"abscor"}.} \item{init.rand}{logical. If \code{init.rand=TRUE}, random number generators are initialized at child processes. Random seeds can be set by \code{seed} argument.} \item{seed}{integer vector of random seeds. It should have the same length as \code{cl}. If \code{NULL} is specified, \code{1:length(cl)} is used as seed vector. The default is \code{NULL}.} } \details{ Function \code{pvclust} conducts multiscale bootstrap resampling to calculate \eqn{p}-values for each cluster in the result of hierarchical clustering. \code{parPvclust} is the parallel version of this procedure which depends on \pkg{snow} package for parallel computation. For data expressed as \eqn{(n \times p)}{(n, p)} matrix or data frame, we assume that the data is \eqn{n} observations of \eqn{p} objects, which are to be clustered. The \eqn{i}'th row vector corresponds to the \eqn{i}'th observation of these objects and the \eqn{j}'th column vector corresponds to a sample of \eqn{j}'th object with size \eqn{n}. There are several methods to measure the dissimilarities between objects. For data matrix \eqn{X=\{x_{ij}\}}{X}, \code{"correlation"} method takes \deqn{ 1 - \frac{ \sum_{i=1}^n (x_{ij} - \bar{x}_j) (x_{ik} - \bar{x}_k) } { \sqrt{\sum_{i=1}^n (x_{ij} - \bar{x}_j)^2} \sqrt{\sum_{i=1}^n (x_{ik} - \bar{x}_k)^2} } }{% 1 - cor(X)[j,k] } for dissimilarity between \eqn{j}'th and \eqn{k}'th object, where \eqn{\bar{x}_j = \frac{1}{n} \sum_{i=1}^n x_{ij} \mbox{and} \bar{x}_k = \frac{1}{n} \sum_{i=1}^n x_{ik}}{cor is function \code{cor}}. \code{"uncentered"} takes uncentered sample correlation \deqn{ 1 - \frac{ \sum_{i=1}^n x_{ij} x_{ik} } { \sqrt{\sum_{i=1}^n x_{ij}^2} \sqrt{\sum_{i=1}^n x_{ik}^2} } }{% 1 - sum(x[,j] * x[,k]) / (sqrt(sum(x[,j]^2)) * sqrt(sum(x[,k]^2))) } and \code{"abscor"} takes the absolute value of sample correlation \deqn{ 1 - \ \Biggl| \frac{ \sum_{i=1}^n (x_{ij} - \bar{x}_j) (x_{ik} - \bar{x}_k) } { \sqrt{\sum_{i=1}^n (x_{ij} - \bar{x}_j)^2} \sqrt{\sum_{i=1}^n (x_{ik} - \bar{x}_k)^2} } \Biggl|. }{% 1 - abs(cor(X)[j,k]). } } \value{ \item{hclust}{hierarchical clustering for original data generated by function \code{hclust}. See \code{\link[stats]{hclust}} for details.} \item{edges}{data frame object which contains \eqn{p}-values and supporting informations such as standard errors.} \item{count}{data frame object which contains primitive information about the result of multiscale bootstrap resampling.} \item{msfit}{list whose elements are results of curve fitting for multiscale bootstrap resampling, of class \code{msfit}. See \code{\link{msfit}} for details.} \item{nboot}{numeric vector of number of bootstrap replications.} \item{r}{numeric vector of the relative sample size for bootstrap replications.} \item{store}{list contains bootstrap replications if \code{store=TRUE} was given for function \code{pvclust} or \code{parPvclust}.} } \seealso{\code{\link{lines.pvclust}}, \code{\link{print.pvclust}}, \code{\link{msfit}}, \code{\link{plot.pvclust}}, \code{\link{text.pvclust}}, \code{\link{pvrect}} and \code{\link{pvpick}}.} \references{ Shimodaira, H. (2004) "Approximately unbiased tests of regions using multistep-multiscale bootstrap resampling", \emph{Annals of Statistics}, 32, 2616-2641. Shimodaira, H. (2002) "An approximately unbiased test of phylogenetic tree selection", \emph{Systematic Biology}, 51, 492-508. Suzuki, R. and Shimodaira, H. (2004) "An application of multiscale bootstrap resampling to hierarchical clustering of microarray data: How accurate are these clusters?", \emph{The Fifteenth International Conference on Genome Informatics 2004}, P034. \url{http://www.is.titech.ac.jp/~shimo/prog/pvclust/} } \examples{ ## using Boston data in package MASS library(MASS) data(Boston) ## multiscale bootstrap resampling boston.pv <- pvclust(Boston, nboot=100) ## CAUTION: nboot=100 may be too small for actual use. ## We suggest nboot=1000 or larger. ## plot/print functions will be useful for diagnostics. ## plot dendrogram with p-values plot(boston.pv) ask.bak <- par()$ask par(ask=TRUE) ## highlight clusters with high au p-values pvrect(boston.pv) ## print the result of multiscale bootstrap resampling print(boston.pv, digits=3) ## plot diagnostic for curve fitting msplot(boston.pv, edges=c(2,4,6,7)) par(ask=ask.bak) ## Print clusters with high p-values boston.pp <- pvpick(boston.pv) boston.pp \dontrun{ ## parallel computation via snow package library(snow) cl <- makeCluster(10, type="MPI") ## parallel version of pvclust boston.pv <- parPvclust(cl, Boston, nboot=1000) } } \author{Ryota Suzuki \email{ryota.suzuki@is.titech.ac.jp}} \keyword{cluster}pvclust/man/pvpick.Rd0000755000175100001440000000427610207330242014333 0ustar hornikusers\name{pvpick} \alias{pvpick} \alias{pvrect} \alias{lines.pvclust} \title{Find Clusters with High/Low P-values} \description{find clusters with relatively high/low \eqn{p}-values. \code{pvrect} and \code{lines} (S3 method for class \code{pvclust}) highlight such clusters in existing plot, and \code{pvpick} returns a list of such clusters. } \usage{ pvpick(x, alpha=0.95, pv="au", type="geq", max.only=TRUE) pvrect(x, alpha=0.95, pv="au", type="geq", max.only=TRUE, border=2, ...) \method{lines}{pvclust}(x, alpha=0.95, pv="au", type="geq", col=2, lwd=2, ...) } \arguments{ \item{x}{object of class \code{pvclust}.} \item{alpha}{threshold value for \eqn{p}-values.} \item{pv}{character string which specifies the \eqn{p}-value to be used. It should be either of \code{"au"} or \code{"bp"}, corresponding to AU \eqn{p}-value or BP value, respectively. See \code{plot.pvclust} for details.} \item{type}{one of \code{"geq"}, \code{"leq"}, \code{"gt"} or \code{"lt"}. If \code{"geq"} is specified, clusters with \eqn{p}-value \emph{greater than or equals} the threshold given by \code{"alpha"} are returned or displayed. Likewise \code{"leq"} stands for \emph{lower than or equals}, \code{"gt"} for \emph{greater than} and \code{"lt"} for \emph{lower than} the threshold value. The default is \code{"geq"}.} \item{max.only}{logical. If some of clusters with high/low \eqn{p}-values have inclusion relation, only the largest cluster is returned (or displayed) when \code{max.only=TRUE}.} \item{border}{numeric value which specifies the color of borders of rectangles.} \item{col}{numeric value which specifies the color of lines.} \item{lwd}{numeric value which specifies the width of lines.} \item{...}{other graphic parameters to be used.} } \value{\code{pvpick} returns a list which contains the following values. \item{clusters}{a list of character string vectors. Each vector corresponds to the names of objects in each cluster.} \item{edges}{numeric vector of edge numbers. The \eqn{i}'th element (number) corresponds to the \eqn{i}'th name vector in \code{clusters}.} } \author{Ryota Suzuki \email{ryota.suzuki@is.titech.ac.jp}} \keyword{aplot}pvclust/man/seplot.Rd0000755000175100001440000000153610366162556014362 0ustar hornikusers\name{seplot} \alias{seplot} \title{Diagnostic Plot for Standard Error of p-value} \description{draws diagnostic plot for standard error of \eqn{p}-value for pvclust object. } \usage{ seplot(object, type=c("au", "bp"), identify=FALSE, main=NULL, xlab=NULL, ylab=NULL, ...) } \arguments{ \item{object}{object of class \code{pvclust}.} \item{type}{the type of \eqn{p}-value to be plotted, one of \code{"au"} or \code{"bp"}.} \item{identify}{logical. If \code{TRUE}, edge numbers can be identified interactively. See \code{\link{identify}} for basic usage.} \item{main, xlab, ylab}{generic graphic parameters. See \code{\link[graphics]{par}} for details.} \item{...}{other graphical parameters to be passed to generic \code{plot} or \code{identify} function.} } \author{Ryota Suzuki \email{ryota.suzuki@is.titech.ac.jp}} \keyword{hplot} pvclust/NAMESPACE0000755000175100001440000000051210366156370013217 0ustar hornikusersexport(pvclust) export(pvpick) export(msfit) export(msplot) export(pvrect) export(seplot) export(parPvclust) S3method(plot, pvclust) S3method(lines, pvclust) S3method(text, pvclust) S3method(print, pvclust) S3method(summary, pvclust) S3method(plot, msfit) S3method(lines, msfit) S3method(summary, msfit) pvclust/R/0000755000175100001440000000000011551334543012175 5ustar hornikuserspvclust/R/pvclust-internal.R0000755000175100001440000001571611551334515015646 0ustar hornikusershc2axes <- function(x) { A <- x$merge # (n,n-1) matrix n <- nrow(A) + 1 x.axis <- c() y.axis <- x$height x.tmp <- rep(0,2) zz <- match(1:length(x$order),x$order) for(i in 1:(n-1)) { ai <- A[i,1] if(ai < 0) x.tmp[1] <- zz[-ai] else x.tmp[1] <- x.axis[ai] ai <- A[i,2] if(ai < 0) x.tmp[2] <- zz[-ai] else x.tmp[2] <- x.axis[ai] x.axis[i] <- mean(x.tmp) } return(data.frame(x.axis=x.axis,y.axis=y.axis)) } hc2split <- function(x) { A <- x$merge # (n-1,n) matrix n <- nrow(A) + 1 B <- list() for(i in 1:(n-1)){ ai <- A[i,1] if(ai < 0) B[[i]] <- -ai else B[[i]] <- B[[ai]] ai <- A[i,2] if(ai < 0) B[[i]] <- sort(c(B[[i]],-ai)) else B[[i]] <- sort(c(B[[i]],B[[ai]])) } CC <- matrix(rep(0,n*(n-1)),nrow=(n-1),ncol=n) for(i in 1:(n-1)){ bi <- B[[i]] m <- length(bi) for(j in 1:m) CC[i,bi[j]] <- 1 } split <- list(pattern=apply(CC,1,paste,collapse=""), member=B) return(split) } pvclust.node <- function(x, r,...) { # require(pvclust) mboot.node <- lapply(r, boot.hclust, nboot=x, ...) return(mboot.node) } boot.hclust <- function(r, data, object.hclust, method.dist, use.cor, method.hclust, nboot, store, weight=F) { n <- nrow(data) size <- round(n*r, digits=0) if(size == 0) stop("invalid scale parameter(r)") r <- size/n pattern <- hc2split(object.hclust)$pattern edges.cnt <- table(factor(pattern)) - table(factor(pattern)) st <- list() # bootstrap start rp <- as.character(round(r,digits=2)); if(r == 1) rp <- paste(rp,".0",sep="") cat(paste("Bootstrap (r = ", rp, ")... ", sep="")) w0 <- rep(1,n) # equal weight na.flag <- 0 for(i in 1:nboot){ if(weight && r>10) { ## <- this part should be improved w1 <- as.vector(rmultinom(1,size,w0)) # resampled weight suppressWarnings(distance <- distw.pvclust(data,w1,method=method.dist,use.cor=use.cor)) } else { smpl <- sample(1:n, size, replace=TRUE) suppressWarnings(distance <- dist.pvclust(data[smpl,],method=method.dist,use.cor=use.cor)) } if(all(is.finite(distance))) { # check if distance is valid x.hclust <- hclust(distance,method=method.hclust) pattern.i <- hc2split(x.hclust)$pattern # split edges.cnt <- edges.cnt + table(factor(pattern.i, levels=pattern)) } else { x.hclust <- NULL na.flag <- 1 } if(store) st[[i]] <- x.hclust } cat("Done.\n") # bootstrap done if(na.flag == 1) warning(paste("inappropriate distance matrices are omitted in computation: r = ", r), call.=FALSE) boot <- list(edges.cnt=edges.cnt, method.dist=method.dist, use.cor=use.cor, method.hclust=method.hclust, nboot=nboot, size=size, r=r, store=st) class(boot) <- "boot.hclust" return(boot) } pvclust.merge <- function(data, object.hclust, mboot){ pattern <- hc2split(object.hclust)$pattern r <- unlist(lapply(mboot,"[[","r")) nboot <- unlist(lapply(mboot,"[[","nboot")) store <- lapply(mboot,"[[", "store") rl <- length(mboot) ne <- length(pattern) edges.bp <- edges.cnt <- data.frame(matrix(rep(0,ne*rl),nrow=ne,ncol=rl)) row.names(edges.bp) <- pattern names(edges.cnt) <- paste("r", 1:rl, sep="") for(j in 1:rl) { edges.cnt[,j] <- as.vector(mboot[[j]]$edges.cnt) edges.bp[,j] <- edges.cnt[,j] / nboot[j] } ms.fitted <- lapply(as.list(1:ne), function(x, edges.bp, r, nboot){ msfit(as.vector(t(edges.bp[x,])), r, nboot)}, edges.bp, r, nboot) class(ms.fitted) <- "mslist" p <- lapply(ms.fitted,"[[","p") se <- lapply(ms.fitted,"[[","se") coef <- lapply(ms.fitted,"[[","coef") au <- unlist(lapply(p,"[[","au")) bp <- unlist(lapply(p,"[[","bp")) se.au <- unlist(lapply(se,"[[","au")) se.bp <- unlist(lapply(se,"[[","bp")) v <- unlist(lapply(coef,"[[","v")) cc <- unlist(lapply(coef,"[[","c")) pchi <- unlist(lapply(ms.fitted,"[[","pchi")) edges.pv <- data.frame(au=au, bp=bp, se.au=se.au, se.bp=se.bp, v=v, c=cc, pchi=pchi) row.names(edges.pv) <- row.names(edges.cnt) <- 1:ne result <- list(hclust=object.hclust, edges=edges.pv, count=edges.cnt, msfit=ms.fitted, nboot=nboot, r=r, store=store) class(result) <- "pvclust" return(result) } dist.pvclust <- function(x, method="euclidean", use.cor="pairwise.complete.obs") { if(!is.na(pmatch(method,"correlation"))){ res <- as.dist(1 - cor(x, method="pearson", use=use.cor)) attr(res,"method") <- "correlation" return(res) } else if(!is.na(pmatch(method,"abscor"))){ res <- as.dist(1 - abs(cor(x,method="pearson",use=use.cor))) attr(res,"method") <- "abscor" return(res) } else if(!is.na(pmatch(method,"uncentered"))){ if(sum(is.na(x)) > 0){ x <- na.omit(x) warning("Rows including NAs were omitted") } x <- as.matrix(x) P <- crossprod(x) qq <- matrix(diag(P),ncol=ncol(P)) Q <- sqrt(crossprod(qq)) res <- as.dist(1 - P/Q) attr(res,"method") <- "uncentered" return(res) } else dist(t(x),method) } corw <- function(x,w, use=c("all.obs","complete.obs","pairwise.complete.obs") ) { if(is.data.frame(x)) x <- as.matrix(x) x <- x[w>0,,drop=F] w <- w[w>0] n <- nrow(x) # sample size m <- ncol(x) # number of variables if(missing(w)) w <- rep(1,n) r <- matrix(0,m,m,dimnames=list(colnames(x),colnames(x))) diag(r) <- 1 use <- match.arg(use) pairu <- F if(use=="all.obs") { u <- rep(T,n) } else if(use=="complete.obs") { u <- apply(x,1,function(y) !any(is.na(y))) } else if(use=="pairwise.complete.obs") { pairu <- T ux <- is.finite(x) } else stop("unknown use") for(i in 1+seq(length=m-1)) { for(j in seq(length=i-1)) { if(pairu) u <- ux[,i] & ux[,j] wu <- w[u]; xi <- x[u,i]; xj <- x[u,j] ws <- sum(wu) if(ws > 1e-8) { xi <- xi - sum(wu*xi)/ws xj <- xj - sum(wu*xj)/ws vxi <- sum(wu*xi*xi)/ws vxj <- sum(wu*xj*xj)/ws if(min(vxi,vxj) > 1e-8) { vxij <- sum(wu*xi*xj)/ws rij <- vxij/sqrt(vxi*vxj) } else { rij <- 0 } } else { rij <- 0 } r[i,j] <- r[j,i] <- rij } } r } ### calculate distance by weight distw.pvclust <- function(x,w,method="correlation", use.cor="pairwise.complete.obs") { if(!is.na(pmatch(method,"correlation"))){ res <- as.dist(1 - corw(x,w, use=use.cor)) attr(res,"method") <- "correlation" return(res) } else if(!is.na(pmatch(method,"abscor"))){ res <- as.dist(1 - abs(corw(x,w, use=use.cor))) attr(res,"method") <- "abscor" return(res) } stop("wrong method") } pvclust/R/pvclust.R0000755000175100001440000003311411551334543014025 0ustar hornikuserspvclust <- function(data, method.hclust="average", method.dist="correlation", use.cor="pairwise.complete.obs", nboot=1000, r=seq(.5,1.4,by=.1), store=FALSE, weight=FALSE) { # data: (n,p) matrix, n-samples, p-variables n <- nrow(data); p <- ncol(data) # hclust for original data METHODS <- c("ward", "single", "complete", "average", "mcquitty", "median", "centroid") method.hclust <- METHODS[pmatch(method.hclust, METHODS)] distance <- dist.pvclust(data, method=method.dist, use.cor=use.cor) data.hclust <- hclust(distance, method=method.hclust) # multiscale bootstrap size <- floor(n*r) rl <- length(size) if(rl == 1) { if(r != 1.0) warning("Relative sample size r is set to 1.0. AU p-values are not calculated\n") r <- list(1.0) } else r <- as.list(size/n) mboot <- lapply(r, boot.hclust, data=data, object.hclust=data.hclust, nboot=nboot, method.dist=method.dist, use.cor=use.cor, method.hclust=method.hclust, store=store, weight=weight) result <- pvclust.merge(data=data, object.hclust=data.hclust, mboot=mboot) return(result) } plot.pvclust <- function(x, print.pv=TRUE, print.num=TRUE, float=0.01, col.pv=c(2,3,8), cex.pv=0.8, font.pv=NULL, col=NULL, cex=NULL, font=NULL, lty=NULL, lwd=NULL, main=NULL, sub=NULL, xlab=NULL, ...) { if(is.null(main)) main="Cluster dendrogram with AU/BP values (%)" if(is.null(sub)) sub=paste("Cluster method: ", x$hclust$method, sep="") if(is.null(xlab)) xlab=paste("Distance: ", x$hclust$dist.method) plot(x$hclust, main=main, sub=sub, xlab=xlab, col=col, cex=cex, font=font, lty=lty, lwd=lwd, ...) if(print.pv) text(x, col=col.pv, cex=cex.pv, font=font.pv, float=float, print.num=print.num) } text.pvclust <- function(x, col=c(2,3,8), print.num=TRUE, float=0.01, cex=NULL, font=NULL, ...) { axes <- hc2axes(x$hclust) usr <- par()$usr; wid <- usr[4] - usr[3] au <- as.character(round(x$edges[,"au"]*100)) bp <- as.character(round(x$edges[,"bp"]*100)) rn <- as.character(row.names(x$edges)) au[length(au)] <- "au" bp[length(bp)] <- "bp" rn[length(rn)] <- "edge #" a <- text(x=axes[,1], y=axes[,2] + float * wid, au, col=col[1], pos=2, offset=.3, cex=cex, font=font) a <- text(x=axes[,1], y=axes[,2] + float * wid, bp, col=col[2], pos=4, offset=.3, cex=cex, font=font) if(print.num) a <- text(x=axes[,1], y=axes[,2], rn, col=col[3], pos=1, offset=.3, cex=cex, font=font) } print.pvclust <- function(x, which=NULL, digits=3, ...) { if(is.null(which)) which <- 1:nrow(x$edges) cat("\n") cat(paste("Cluster method: ", x$hclust$method, "\n", sep="")) cat(paste("Distance : ", x$hclust$dist.method, "\n\n", sep="")) cat("Estimates on edges:\n\n") print(round(x$edges[which,], digits=digits)) cat("\n") } summary.pvclust <- function(object, ...){ class(object) <- "list" summary(object, ...) } pvrect <- function(x, alpha=0.95, pv="au", type="geq", max.only=TRUE, border=2, ...) { len <- nrow(x$edges) member <- hc2split(x$hclust)$member order <- x$hclust$order usr <- par("usr") xwd <- usr[2] - usr[1] ywd <- usr[4] - usr[3] cin <- par()$cin ht <- c() j <- 1 if(is.na(pm <- pmatch(type, c("geq", "leq", "gt", "lt")))) stop("Invalid type argument: see help(pvrect)") for(i in (len - 1):1) { if (pm==1) wh <- (x$edges[i,pv] >= alpha) # Greater than or EQuals else if(pm==2) wh <- (x$edges[i,pv] <= alpha) # Lower than or EQuals else if(pm==3) wh <- (x$edges[i,pv] > alpha) # Greater Than else if(pm==4) wh <- (x$edges[i,pv] > alpha) # Lower Than if(wh) { mi <- member[[i]] ma <- match(mi, order) if(max.only == FALSE || (max.only && sum(match(ma, ht, nomatch=0)) == 0)) { xl <- min(ma) xr <- max(ma) yt <- x$hclust$height[i] yb <- usr[3] mx <- xwd / length(member) / 3 my <- ywd / 200 rect(xl - mx, yb + my, xr + mx, yt + my, border=border, shade=NULL, ...) j <- j + 1 } ht <- c(ht, ma) } } } msplot <- function(x, edges=NULL, ...) { if(is.null(edges)) edges <- 1:length(x$msfit) d <- length(edges) mfrow.bak <- par()$mfrow on.exit(par(mfrow=mfrow.bak)) par(mfrow=n2mfrow(d)) for(i in edges) { if(i == 1 || (i %% 10 == 1 && i > 20)) main <- paste(i, "st edge", sep="") else if(i == 2 || (i %% 10 == 2 && i > 20)) main <- paste(i, "nd edge", sep="") else if(i == 3 || (i %% 10 == 3 && i > 20)) main <- paste(i, "rd edge", sep="") else main <- paste(i, "th edge", sep="") plot(x$msfit[[i]], main=main, ...) } } lines.pvclust <- function(x, alpha=0.95, pv="au", type="geq", col=2, lwd=2, ...) { len <- nrow(x$edges) member <- hc2split(x$hclust)$member order <- x$hclust$order usr <- par("usr") xwd <- usr[2] - usr[1] ywd <- usr[4] - usr[3] cin <- par()$cin ht <- c() j <- 1 if(is.na(pm <- pmatch(type, c("geq", "leq", "gt", "lt")))) stop("Invalid type argument: see help(lines.pvclust)") for(i in (len - 1):1) { if (pm==1) wh <- (x$edges[i,pv] >= alpha) # Greater than or EQuals else if(pm==2) wh <- (x$edges[i,pv] <= alpha) # Lower than or EQuals else if(pm==3) wh <- (x$edges[i,pv] > alpha) # Greater Than else if(pm==4) wh <- (x$edges[i,pv] > alpha) # Lower Than if(wh) { mi <- member[[i]] ma <- match(mi, order) if(sum(match(ma, ht, nomatch=0)) == 0) { xl <- min(ma) xr <- max(ma) yt <- x$hclust$height[i] yb <- usr[3] mx <- xwd/length(member)/10 segments(xl-mx, yb, xr+mx, yb, xpd=TRUE, col=col, lwd=lwd, ...) j <- j + 1 } ht <- c(ht, ma) } } } pvpick <- function(x, alpha=0.95, pv="au", type="geq", max.only=TRUE) { len <- nrow(x$edges) member <- hc2split(x$hclust)$member order <- x$hclust$order ht <- c() a <- list(clusters=list(), edges=c()); j <- 1 if(is.na(pm <- pmatch(type, c("geq", "leq", "gt", "lt")))) stop("Invalid type argument: see help(pickup)") for(i in (len - 1):1) { if (pm==1) wh <- (x$edges[i,pv] >= alpha) # Greater than or Equals else if(pm==2) wh <- (x$edges[i,pv] <= alpha) # Lower than or Equals else if(pm==3) wh <- (x$edges[i,pv] > alpha) # Greater Than else if(pm==4) wh <- (x$edges[i,pv] > alpha) # Lower Than if(wh) { mi <- member[[i]] ma <- match(mi, order) if(max.only == FALSE || (max.only && sum(match(ma, ht, nomatch=0)) == 0)) { a$clusters[[j]] <- x$hclust$labels[mi] a$edges <- c(a$edges,i) j <- j + 1 } ht <- c(ht, ma) } } a$edges <- a$edges[length(a$edges):1] a$clusters <- a$clusters[length(a$edges):1] return(a) } parPvclust <- function(cl, data, method.hclust="average", method.dist="correlation", use.cor="pairwise.complete.obs", nboot=1000, r=seq(.5,1.4,by=.1), store=FALSE, weight=FALSE, init.rand=TRUE, seed=NULL) { if(!(require(snow))) stop("Package snow is required for parPvclust.") if((ncl <- length(cl)) < 2 || ncl > nboot) { warning("Too small value for nboot: non-parallel version is executed.") return(pvclust(data,method.hclust,method.dist,use.cor,nboot,r,store)) } if(init.rand) { if(is.null(seed)) seed <- 1:length(cl) else if(length(seed) != length(cl)) stop("seed and cl should have the same length.") # setting random seeds parLapply(cl, as.list(seed), set.seed) } # data: (n,p) matrix, n-samples, p-variables n <- nrow(data); p <- ncol(data) # hclust for original data METHODS <- c("ward", "single", "complete", "average", "mcquitty", "median", "centroid") method.hclust <- METHODS[pmatch(method.hclust, METHODS)] distance <- dist.pvclust(data, method=method.dist, use.cor=use.cor) data.hclust <- hclust(distance, method=method.hclust) # multiscale bootstrap size <- floor(n*r) rl <- length(size) if(rl == 1) { if(r != 1.0) warning("Relative sample size r is set to 1.0. AU p-values are not calculated\n") r <- list(1.0) } else r <- as.list(size/n) nbl <- as.list(rep(nboot %/% ncl,times=ncl)) if((rem <- nboot %% ncl) > 0) nbl[1:rem] <- lapply(nbl[1:rem], "+", 1) cat("Multiscale bootstrap... ") mlist <- parLapply(cl, nbl, pvclust.node, r=r, data=data, object.hclust=data.hclust, method.dist=method.dist, use.cor=use.cor, method.hclust=method.hclust, store=store, weight=weight) cat("Done.\n") mboot <- mlist[[1]] for(i in 2:ncl) { for(j in 1:rl) { mboot[[j]]$edges.cnt <- mboot[[j]]$edges.cnt + mlist[[i]][[j]]$edges.cnt mboot[[j]]$nboot <- mboot[[j]]$nboot + mlist[[i]][[j]]$nboot mboot[[j]]$store <- c(mboot[[j]]$store, mlist[[i]][[j]]$store) } } result <- pvclust.merge( data=data, object.hclust=data.hclust, mboot=mboot) return(result) } msfit <- function(bp, r, nboot) { if(length(bp) != length(r)) stop("bp and r should have the same length") nboot <- rep(nboot, length=length(bp)) use <- bp > 0 & bp < 1 p <- se <- c(0,0); names(p) <- names(se) <- c("au", "bp") coef <- c(0,0); names(coef) <- c("v", "c") a <- list(p=p, se=se, coef=coef, df=0, rss=0, pchi=0); class(a) <- "msfit" if(sum(use) < 2) { # if(mean(bp) < .5) a$p[] <- c(0, 0) else a$p[] <- c(1, 1) if(mean(bp) < .5) a$p[] <- c(0, bp[r==1.0]) else a$p[] <- c(1, bp[r==1.0]) return(a) } bp <- bp[use]; r <- r[use]; nboot <- nboot[use] zz <- -qnorm(bp) vv <- ((1 - bp) * bp) / (dnorm(zz)^2 * nboot) a$use <- use; a$r <- r; a$zz <- zz X <- cbind(sqrt(r), 1/sqrt(r)); dimnames(X) <- list(NULL, c("v","c")) fit <- lsfit(X, zz, 1/vv, intercept=FALSE) a$coef <- coef <- fit$coef h.au <- c(1, -1); h.bp <- c(1, 1) z.au <- drop(h.au %*% coef); z.bp <- drop(h.bp %*% coef) a$p["au"] <- pnorm(-z.au); a$p["bp"] <- pnorm(-z.bp) V <- solve(crossprod(X, X/vv)) vz.au <- drop(h.au %*% V %*% h.au); vz.bp <- drop(h.bp %*% V %*% h.bp) a$se["au"] <- dnorm(z.au) * sqrt(vz.au); a$se["bp"] <- dnorm(z.bp) * sqrt(vz.bp) a$rss <- sum(fit$residual^2/vv) if((a$df <- sum(use) - 2) > 0) { a$pchi <- pchisq(a$rss, lower.tail=FALSE, df=a$df) } else a$pchi <- 1.0 return(a) } plot.msfit <- function(x, curve=TRUE, main=NULL, sub=NULL, xlab=NULL, ylab=NULL, ...) { if(is.null(main)) main="Curve fitting for multiscale bootstrap resampling" if(is.null(sub)) { sub <- paste("AU = ", round(x$p["au"], digits=2), ", BP = ", round(x$p["bp"], digits=2), ", v = ", round(x$coef["v"], digits=2), ", c = ", round(x$coef["c"], digits=2), ", pchi = ", round(x$pchi, digits=2)) } if(is.null(xlab)) xlab=expression(sqrt(r)) if(is.null(ylab)) ylab=expression(z-value) a <- sqrt(x$r); b <- x$zz if(!is.null(a) && !is.null(b)) { plot(a, b, main=main, sub=sub, xlab=xlab, ylab=ylab, ...) if(curve) lines(x, ...) } else if (!is.null(a)){ plot(0, 0, main=main, sub=sub, xlab=xlab, ylab=ylab, type="n", xaxt="n", yaxt="n", ...) a <- text(mean(a), 0, "No fitting") } } lines.msfit <- function(x, col=2, lty=1, ...) { v <- x$coef["v"]; c <- x$coef["c"] curve(v * x + c / x, add=TRUE, col=col, lty=lty) } summary.msfit <- function(object, digits=3, ...) { cat("\nResult of curve fitting for multiscale bootstrap resampling:\n\n") cat("Estimated p-values:\n") pv <- data.frame(object$p, object$se) names(pv) <- c("Estimate", "Std. Error"); row.names(pv) <- c("au", "bp") print(pv, digits=digits); cat("\n") cat("Estimated coefficients:\n") coef <- object$coef print(coef, digits=digits); cat("\n") cat(paste("Residual sum of squares: ", round(object$rss,digits=digits)), ", p-value: ", round(object$pchi, digits=digits), " on ", object$df, " DF\n\n", sep="") } seplot <- function(object, type=c("au", "bp"), identify=FALSE, main=NULL, xlab=NULL, ylab=NULL, ...) { if(!is.na(pm <- pmatch(type[1], c("au", "bp")))) { wh <- c("au", "bp")[pm] if(is.null(main)) main <- "p-value vs standard error plot" if(is.null(xlab)) xlab <- c("AU p-value", "BP value")[pm] if(is.null(ylab)) ylab <- "Standard Error" plot(object$edges[,wh], object$edges[,paste("se", wh, sep=".")], main=main, xlab=xlab, ylab=ylab, ...) if(identify) identify(x=object$edges[,wh], y=object$edges[,paste("se", wh, sep=".")], labels=row.names(object$edges)) } else stop("'type' should be \"au\" or \"bp\".") }