Data compression algorithms were usually designed for data processing symbol by symbol. Symbols are usually characters or bytes, but several other techniques may be used. The most well-known approach is using syllables or words as symbols. Another approach is to take 2-grams, 3-grams or any n-grams as a symbols. All these approaches has pros and cons, but none of them is the best for any file. This paper describes approach of evolving alphabet from characters and 2-grams, which is optimal for compressed text files. The efficiency of the approach will be tested on three compression algorithms.