The easiest way to understand it is to realize firstly that these so-called groups used to identify an admixture are not set in stone, they change with every analysis depending on the needs of the study. If you take West Asian and try to understand it in terms of history and archaeology it will drive you around the bend. Don't think of it as a 'standard' group, it is more helpful to view this group as a 'measuring instrument' used to illustrate differences between populations more clearly.
For example, if we compare Poland and Greece, the admixture known as 'West Asian' will incorporate more South Asian and less West Asian to bring out the difference between Polish and Greek admixtures. Greeks and Polish may have similar levels of West Asian, but in such a comparison Polish should indicate much less West Asian and more North European. Contrast is good and altering one calibrates the measurement to get the best contrast. The names stay the same but their respective values and make-up alter to get the best effect.
Conversely, if we compare Sicilians and Greeks the admixture known as 'West Asian' will be dominant in Sicilians and North European identifiably more in Greeks. Where did they get North European in Greeks all of a sudden you may ask, because it is where the difference lies, the only difference is that now the weight of North European within Greeks is elevated and it appears proportionally higher than it would otherwise seem.
The Dodecad program puts more emphasis on a characteristic unique to a certain population. This characteristic may only be 5% of the population but it is significant if not found in the population you want to compare with. Therefore the proportion of 5% is in fact illustrated as 100% and 2.5% as 50% etc. etc. This is only for effect and helps make the comparison more representative and easier to read. Ultimately these types of analyzes are of limited value when comparing similar populations.