Biology now produces huge amount of data mainly in the form of symbolic sequences. These data are noisy and incomplete so that statistics is inevitable as a first step in any analysis. However, in order to reveal the biological regularities masked by billion years of random mutations and natural selection one must invoke more or less deterministic approaches. We will draw a few simple examples from our recent bioinformatics work that make use of Poisson distribution and Markov prediction in statistics, Goulden-Jackson cluster method in enumerative combinatorics, number of Eulerian loops in graph theory, and factorizable language in formal language theory.