skipping the first 4 rows... rather than use the colon operator (':') by itself... use ('5:end')
The ':' by itself is the same as '1:end', i.e. a(:)==a(1:end)
Reading in the data you will probably have an easier time with the 'textscan' function. You can specify delimiters (';' in your case). As your first few lines aren't marked with a symbol to show that they're comments/heading... you'll probably have to load all the data and then remove the first 4 rows.