The CS difference between predicted and observed

##                          diff
## bmr7256.str  0.01  0.48 -0.47
## bmr4961.str -0.03  0.11 -0.14
## bmr6208.str -0.04  0.37 -0.41
## bmr5686.str  0.45  0.80 -0.35
## bmr7068.str -0.26 -0.03 -0.23
## bmr6717.str  0.03 -0.43  0.46

From the histgram there seems to have two distributions overlap in the “Difference of difference” histagram, so I am trying to use it.

Here I used the mixtool to classified the two distribution

## number of iterations= 47
## [1]  0.2761755 -0.1147990
## [1] 0.9319679 0.2598826
## number of iterations= 461
## [1] -0.1484926 -0.0462678
## [1] 0.2072433 0.3472272

  • Classification 1: without remove the extreme value, mean are -0.1147995 and 0.2761492, and standard deviations are 0.2598810 and 0.9319426.

  • Classification 2: removed all the extreme value and keep the ones [-1,1] mean are -0.1484851 and -0.0462389, and standard deviations are 0.2072632 and 0.3472506. From the above two attemption for classification, it’s really hard to claiming there are two distributions. So the only logical way to make sense of the data is consider the data as one normal distribution, and only use the the samples that betwen +/- 1 standard deviation, which is roughly 67% of the data.

And the statistics are:

## [1] -0.1058591
## [1] 0.2791205

The new dataset has following number of observation.

## [1] 1647
## [1] "The target obs number is 682"
## [1] "The non-target obs number is 965"

Estimate the covariance

  1. There is one issue the data from the dataset and the secondary structures report had diffirent files. So the data is used to estimate a better covarriance may not be in the secondary structure report. Using the intersect.
##       v     p.value               
##  [1,] "A-B" "0.606524212801048"   
##  [2,] "A-C" "0.00470358738007581" 
##  [3,] "A-H" "0.301889034908116"   
##  [4,] "R-B" "0.3352833501262"     
##  [5,] "R-C" "0.0301051010331346"  
##  [6,] "R-H" "9.29264905513705e-06"
##  [7,] "N-B" "0.0178121305915855"  
##  [8,] "N-C" "0.29347131725404"    
##  [9,] "N-H" "0.0136513011448303"  
## [10,] "D-B" "0.289203969489064"   
## [11,] "D-C" "5.39802150774094e-05"
## [12,] "D-H" "1.21333128957346e-05"
## [13,] "C-B" "0.6079557601842"     
## [14,] "C-C" "0.201210238083825"   
## [15,] "C-H" "0.000175969222676153"
## [16,] "Q-B" "0.94033100673002"    
## [17,] "Q-C" "0.0425021556066736"  
## [18,] "Q-H" "0.491614708073569"   
## [19,] "E-B" "0.811637168177298"   
## [20,] "E-C" "6.39472809502717e-05"
## [21,] "E-H" "1.87173949079966e-05"
## [22,] "H-B" "0.075075327818422"   
## [23,] "H-C" "0.948854830368885"   
## [24,] "H-H" "0.864965442525413"   
## [25,] "I-B" "0.694038566249678"   
## [26,] "I-C" "0.630249960631841"   
## [27,] "I-H" "0.0246734825282535"  
## [28,] "L-B" "0.714057940533835"   
## [29,] "L-C" "0.00702429870075738" 
## [30,] "L-H" "0.843592671655967"   
## [31,] "K-B" "0.902762069398099"   
## [32,] "K-C" "0.475682645972612"   
## [33,] "K-H" "0.0806406161309812"  
## [34,] "M-B" "0.754685434976996"   
## [35,] "M-C" "0.0406176811901431"  
## [36,] "M-H" "0.407871070714918"   
## [37,] "F-B" "0.718522194363658"   
## [38,] "F-C" "0.0223532883647537"  
## [39,] "F-H" "0.0714693968163156"  
## [40,] "P-B" "0.898782125835033"   
## [41,] "P-C" "0.253164773984403"   
## [42,] "P-H" "0.922211328270108"   
## [43,] "S-B" "0.208941852122058"   
## [44,] "S-C" "0.512231566956487"   
## [45,] "S-H" "0.103302123492859"   
## [46,] "T-B" "5.31092992028803e-07"
## [47,] "T-C" "2.25393037567301e-11"
## [48,] "T-H" "1.96079819048123e-10"
## [49,] "Y-B" "0.0735292326777213"  
## [50,] "Y-C" "0.0145202458255251"  
## [51,] "Y-H" "0.551235580452909"   
## [52,] "W-B" "0.116513555687378"   
## [53,] "W-C" "0.79805842893556"    
## [54,] "W-H" "0.639951525876334"   
## [55,] "V-B" "1.61818802801861e-09"
## [56,] "V-C" "0.00014477180312622" 
## [57,] "V-H" "0.00351473297351323"

Number of pvalues less than 0.05

sum(p[,2]<0.05)
## [1] 13