Far too complex.
You are trying to define strings which will be magically evaluated and converted back into data. The syntax you invented (passing character vectors/strings to the plot function) simply does not work. Did you see this in the plot documentation? This could be made to work, but it would be a very slow, complex, obfuscated, and buggy way to write code. Read this to know why:
You also throw away most of your strings here anyway:
finalPlot = toPlot{:};
Method One: comma-separate list: there is no reason why you need to use strings at all. Why bother? Just put all of the numeric data into a cell array and then use a comma-separated list when calling plot. First lets define some sample data:
>> N = 7;
>> Y = randi(9,N,5)
Y =
8 7 8 2 5
5 2 7 8 2
6 3 9 9 9
4 9 6 9 8
3 8 1 7 9
2 7 4 4 1
3 6 1 4 5
>> X = linspace(0,1,N);
and then select which data to plot using the logical index idx:
C = num2cell(Y,1);
C = repmat(C,3,1);
C(1,:) = {X};
C(3,:) = {'k','b','c','g','m'};
idx = [true,true,false,false,true];
plot(C{:,idx}) % comma-separated list
Method two: basic indexing: here is an another solution using basic MATLAB indexing and NaN's (which are not plotted). The same columns will always have the same color, regardless of which columns you plot.
>> plot(X(:),Y+(0./idx))
>> legend('L1','L2','L3','L4','L5')
Just change the index vector and you will get only those columns that you request:
>> idx = [false,true,false,true,false];
>> plot(X(:),Y+(0./idx))
>> legend('L1','L2','L3','L4','L5')
To make this work for your situation just need to concatenate those vectors into one matrix (which they should be already):
Y = [plotLine1(:),plotLine2(:),plotLine3(:),plotLine4(:),plotLine5(:)];
and define the index something like this (not that str2num is best avoided):
idx = str2double(a)==1;
You can change the color by setting the axes' line ColorOrder property.
