I'm a bit confused about your explanation. Are you using the transparent pixels to store extra data, or are you using the alpha channel to store extra data? And if you're using the transparent pixels, then what data is it and how do your shaders know how to find the transparent pixels?