Method of data evaluation 
17 
The following example Illustrates this comparison process (the two samples to 
be compared consist of three peaks each (peak 1-3)): 
Step 1 (raw data and ratios) Step 2 (features) Step 3 (conclusion matrix) 
peak 
raw [pA] 
norm** 
1 
193.9 
0.242 
RSD 
CD 
RSD 
CD 
2 
257.6 
0.322 
1 on 1 
5.1% 
10.2% 
33% 
100% 
3 
329.6 
0.412 
2 on 2 
4.8% 
9.6% 
1 
214.7 
0.268 
3 on 3 
5.2% 
10.3% 
2 
283.6 
0.355 
3 
365.5 
0.457 
Criterion 
< 5% 
< 14% 
> 50% 
Normalized on virtual maximum of 800 [pA] 
Step 1: The extracted heights are normalized on a hypothetical maximum of 
800 pA. In order to narrow the number of features, a pre-sorting was con 
ducted: Every peak smaller than 10% of the maximum was not considered In 
following computations. In our example above, all normalized peaks a larger 
than 0.1, so they were all used to calculate features. 
Step 2: Creation of the two features of RSD and CD. A green background Indi 
cates that the result of this feature Is below criterion limits, and therefore statis 
tically similar. If these features exceed their criterion, the background Is 
coloured In red. This means that the compared peaks of the samples are statis 
tically dissimilar. 
Step 3: The amount of features showing matches are compared to the number 
of total features and displayed as ratios. 
The first step of extracting peak Intensities and normalisation to the maximum 
should not need further explanation. 
Based on the normalised (“norm”) peak values of step 1, for every peak com 
parison the RSD- and CD-features are created. Looking at the RSD column 
(step 2) we see that only the RSD of the second peak’s comparison Is below 
5 % (green background). Since this Is only one feature out of three, this results 
In a matching-ratio of 33% (step 3). In contrast, for CD, step 2 shows that all 
three peak comparisons are below the threshold value of 14%, which leads to 
a 100% matching-ratio In step 3 for CD 1 . 
1 Note: the nature and usability of both of these features (RSD and CD) and how to 
proceed and handle data in case of contradicting results, Is not part of this investiga 
tion. Anyway, in the following, both values are presented.