From 0653a8c1e3169f04bc8d8eb52bdb9ac266e28f62 Mon Sep 17 00:00:00 2001
From: Tanushree Tunstall <tanu@tunstall.in>
Date: Tue, 11 Feb 2020 15:02:32 +0000
Subject: [PATCH] updated ref dict to inc aa_calcprop

---
 .../__pycache__/reference_dict.cpython-37.pyc | Bin 2216 -> 2624 bytes
 meta_data_analysis/reference_dict.py          |  21 +++++++++++++-----
 2 files changed, 16 insertions(+), 5 deletions(-)

diff --git a/meta_data_analysis/__pycache__/reference_dict.cpython-37.pyc b/meta_data_analysis/__pycache__/reference_dict.cpython-37.pyc
index 3c936f99b195bc3763249c3d77aedeb5bb30dba3..966f27fd47870a9d02d668089f99fb3eb4ea54f8 100644
GIT binary patch
literal 2624
zcmbtV&2!U66yMc{Ez5T7ge0~D0YeK#e8fpQ(*l#}kdU^t1%?u)p&F=0c$Zj3mc&Yi
z*wg8SOHT}^o^nWz?SIe<f5=`r^}-*(sc%=Y<)oPoGquN0Z-4u~-_!G2xojf%{PpMB
zPxV=Z-ipKUBW|w4FaAwIh#(sgOcWp6ihQcJN;RtP>U-!V#)zsv=XUc%-NUv4Xn`6;
ztD!w{GbTlJ+k~534cSFnB>5VmW(^So&LvW?%fz%NXo;2qp8%`^x)o_%*{#wF*q)@5
zyHljN&mAKr;HF8LOppqxl1VZ}j*;n|ygdWh44DP2M(Thz$Q(JoQ?O@8_z*p@W7st^
zPfn6kJBnQ=O>&x?As>;CcQm^}&XRNF6F8lR)7)#k{R^GmJ-&z5(QEuDc&OMX?0I_f
zC8G1hT0!)b<SvZ3i@>F~mqy&>5x1~{wioHlI(l^8Zi0nY1JUNw+-uAp3VS+&^9*e^
z(C0`+E8y|j!%eikKxaYY+#!uupmCvrc2Cn;p&{xH@~@<P?FjidAb(NH*T&@Efczq@
z9}&?zkiR4&svn5xoz!UjOG9BIqsvmGF{YtBYf6o|BQ)-T#)8zCJD_oIQ{+<JL`2y@
z*5cr*d4o|mp`;lG&HFvt{I(Y~m##FImacued~Nwkb9r&;s%c(#dr2o`*P4kN^x_Vq
z)cjkF_MP|m*TLD=mY1~F-NbE0edZ3l!0md3w!9$fC9RNoTY&hyG=4cKjFjHv)*zP(
z13DBRlv}Rrv_nGUg?9X$*An5?@oAD!CUM6PUr@$%FCcWA>uYZ8wRvu}?GgEe5a3z7
z*(WTFI^knDnDVgQ_klN7S=e<G;PM+iKkD?DmA6!m?{R#S<L@km<8_X|6wf;xe{bbD
z{+i=ej(_0z7RTRkd{^9UaC~2=eZ}!6$3F__2FJI#F5+l&{ch;HY!rnjJnI`>*Y|m0
z)%T*#egqM1d#-O4ER{__Sy@GpDiBLylY*ZTyvB|T{)8aM1eq3Sp49~35Tq{1X+h=%
zIVH$RAXZ^aRT>}ZOVg<n;ZXV?dYTactTM#rgb^*ZDh#Hkg(X0b@DGVJg_yJ;m5u+-
ztuz*37Mrv&$|5UETBWQqrK4gvRD(1inT?D}mQiM~Uz~&4rs7$xr8AM~|0)A7cG7{A
zqcA*(>9iE+<S3)GzoKxm$!Wf#`_@F-5Lp@7WEr;XsT={>#B2^^0<uA(Z5%(4o#Qz-
zeNXhq^0mnaw3jf~9~6QxxEzV0wX_4XryeRk1S><3-m`{f11h>^8ELRG&O-w#bISr#
z1G>f4D2%!OLGJbJ4H%mo>C_yHOt$+2+#9fLB3Kkle8?^H0$_>T_d~|Z``6A3*p-Y|
zhesKzm-KlNkhbf$g$=tB1_w5kb%y61hlFj%;kn2Sh#PYyB3zN#GA~GYZi^=E7ldn(
zn{>EIw}n_Q24DG3*rmjSMc3|#helX$i{}ZfTXD=o1LbuN)~(PjdV%xA^C>K0$^jHS
zH(<#VN9<9(+jm@-7h;+?>E_HQp#!Tj;ikOW=b!al-%C6iFL86kFZ1e%AC{()A!%`@
zGGJI9_L-PH-AiaU=J+Yc&$${Wkc;od3D>t+*o$B!qlgCN55yjTv?TA>S|a<F_}7vD
z7q>QU-oNd*@ORRWy|~q-iR*|;WDuiID5F7}$~+dLJ~vjnA?f+_i#(W(e+Qze1~v*M
zT*4E$q-lzQtEz@II2*dDnguhj;OuK?8J}{7x{oN?d58_A4SiNI@#qWu1fG+$K1Nd{
GrT+j+qi1{o

literal 2216
zcmbtU&2!sC6yMc{B`c2YBu!Gc6x`222kImoXgirqlLjczb_(eb*r8~)-X#{5C1oYI
zb`D(n59Cg!lbjg-6;3_xl@tF$PwZR8R+7nZ!5Y7P``h>4zOQ{7)6*VOzu*73^7HjM
zg#L!j_y;HV)jzWggcu@-VP-^_7<x5{$t`Xk+DGUO#)zB0Rt_Cz9%14tw8~v(Euteh
zIVoivddyZbQ+(>M#wy3$G;<VJXD+KUk4>=}n`U*^IB-ZqVNEupuvs>z&<kvyEgV!y
zbApf2#RHeju#2q4E*%(TmR)8auq*6C_R)bw=JxS$jn@w^9HC9LkDo97WRQ8Xz!%>j
zzQEe6h+ov))d_b^amDHN33p=^4VQRh6FvW!w3Ob>7UHd!m3=H~V8upi-{7ql`V5(9
zRo$~&+eX9Nys0#9ozYlR8dqEB@Dgu=2INV}uW9+2bL3+se@)BJoRE)|{1Tr%$D_AO
z{<`*P_LN6&wZ_~z8t;_G4XrVEO5@!&+Wl)AF=Gq)xAyLO4+N(<XYC|zKO6A&*Mqpd
ze5bv<eDBkhdn<R^D@)6FJ?}mp<h?}PYv(i`WIe&T_YbtyllS<yy_=n#Fz;;AoOaTY
zpnG9V`yu0<Fir<~ClTR}LZZDIr9n4gJiFb^UXjHdtOOCyb1pO<M9FI|q#eeL4~e}&
zv#?7l>)nv)Rhd-l)WZ=IN!m+ZsEwyL!%?JocU>fXnk&xP8boPtAbiI+CH_w0hZ29|
z8xn6y{5h;oB>vW~Nc^S5>k@x2@gs@9lK3f{ZAttL)V`2-TjC!9dLZ#*VuK%DVn0nH
zDkffpaI?MDr%^<z>rt5Yzys-QrMw%`$gldQuvD6a4Wy!oZwLo?7kEp|06z<)3ItxL
z?}%yO8$jwn7J;b8jKn;U3yS#F6RO(yL|>aOE`ba6{o|Xa0H^_Jf)T5*3I+?)U`e6p
z+^+&z05NSrE4%+4TkV8{rEl8CB#JUEMV89U)XwwaOb!Z%bTry4okktOadJ*aTZosr
z)|Ww~|0@k-s#i3mZiVp%y3;pQB_|OT^@YUgA{X&O_5FHL5S<xaWbL-Dscr#X#Ig@`
z0J=a>Hi>`Gl_M3Z--)zez_XD~%m=wpkFFXg@h2%Xt#6&`J@ZWRm9u4m%Yu~+3RUyn
z!ddO-nFiIdO&lYxdhMxtsml{?g;W)m(@~TNGJSj;yjH_Ukmh)+T@CXQX(aKf>EK_M
zP7tu98w9eF(wNaq8Yz=Tl1V!iVVuj#4$r%<nY2=x_oT;%AU4R<OmTWipR-U6mu?Rh
zSIs;Ys4*n%{wSbSRx_RldhkdmPXaY*a_Q+K%@c2TK%+1ZdA59lUm?v2J`Su=B1LGk
zB`^+7&A6QRGl^eH{7RZxE~FKOSuX7zkqlB<p=ruv_9qf6&@e#WZ*(Ar4$NOYZXazu
zeD*k?YOs#7FzfVrP6MV!a0v<i3l}`@avj3$bR^x?e!>P3UxT*HhKg9Gi`{As*Ki%z
pEYI|+o@3zhbFGpuIM+NzjB-E5TzkRr@bRa(I=-ao#S^qZ{ski>^j!b|

diff --git a/meta_data_analysis/reference_dict.py b/meta_data_analysis/reference_dict.py
index e83ce2c..b274df2 100755
--- a/meta_data_analysis/reference_dict.py
+++ b/meta_data_analysis/reference_dict.py
@@ -13,20 +13,19 @@ import os
 
 #!#########################!
 # REQUIREMNETS:
-# Data_original/ must exist
+# Data/ must exist
 # containing GWAS data
 #!#########################!
 
-#print(os.getcwd()) 
-#homedir = os.path.expanduser('~') # spyder/python doesn't recognise tilde
-#os.chdir(homedir + '/git/Data/pyrazinamide/input/original') 
+print(os.getcwd()) 
+homedir = os.path.expanduser('~') # spyder/python doesn't recognise tilde
+os.chdir(homedir + '/git/Data/pyrazinamide/input/original') 
 print(os.getcwd())
 
 #%%
 ############# specify variables for input and output paths and filenames
 drug = 'pyrazinamide'
 #gene = 'pnca'
-
 datadir = homedir + '/git/Data'
 basedir =  datadir + '/' + drug + '/input'
 
@@ -100,6 +99,13 @@ qualities_polarity = { ('D', 'E'): 'acidic'
                       , ('A', 'F', 'I', 'L', 'M', 'P', 'V', 'W'): 'non-polar'    
 }
 
+# almost same as the one above
+aa_calcprop = { ('D', 'E'): 'neg'
+                      , ('H', 'K', 'R'): 'pos'
+                      , ('N', 'Q', 'S', 'T', 'Y'): 'polar'
+                      , ('C', 'G', 'A', 'F', 'I', 'L', 'M', 'P', 'V', 'W'): 'non-polar'    
+}
+
 #==============================================================================                
 #adding amino acid properties to my dict of dicts                      
 for k, v in my_aa_dict.items():
@@ -109,6 +115,7 @@ for k, v in my_aa_dict.items():
     v['aa_taylor'] = list() #initialise keys (allows for overalpping properties)
     v['aa_prop_water'] = str() #initialise keys
     v['aa_prop_polarity'] = str() #initialise keys
+    v['aa_calcprop'] = str() #initialise keys
     
     for group in qualities1:
         if v['one_letter_code'] in group:
@@ -129,6 +136,10 @@ for k, v in my_aa_dict.items():
     for group in qualities_polarity:
         if v['one_letter_code'] in group:
             v['aa_prop_polarity']+= qualities_polarity[group] # += for str concat 
+            
+    for group in aa_calcprop:
+        if v['one_letter_code'] in group:
+            v['aa_calcprop']+= aa_calcprop[group] # += for str concat 
              
 #COMMENT:VOILA!!! my_aa_dict is now a dict of dicts containing all associated properties for each aa
 #==============================================================================