calc_sscore: change normalization technique

[Paul Villanueva]

MetaFunPrimer/src/calc_sscore

Lines 8 to 22 in d65da5d

 info=$(awk '{ 
     for(i = 2; i <= NF; i++) 
     {  
         sums[i] += $i 
         sums_squared[i] += ($i)^2 
     } 
 } 
 END { 
     for(i = 2; i <= NF; i++) 
     {  
         means[i - 1] = sums[i] / NR; 
         std_devs[i - 1]=sqrt((sums_squared[i]-sums[i]^2/NR)/NR) 
     } 
     printf "%s,%s,%s,%s", means[1],std_devs[1],means[2],std_devs[2]  
 }' $input_file) 

The S-score is currently calculated by taking the mean of the standardized presence and absences of a gene. However, this isn't compatible with our desired thresholding function, since the the first gene will always have S-score greater than whatever threshold that is set. Change the above lines so that the S-score is the mean of the 0-1 normalized presence and abundance.

[Paul Villanueva]

Resolved by #21