The Sciences are experiencing an unprecedented explosion in the amount of 
available data. Traditional data analysis methods cannot deal with this 
increased quantity of data.  There is an urgent need to automate the process
of refining scientific data into scientific knowledge. Inductive logic 
programming (ILP) is a data analysis framework well suited for this task. 
We have applied ILP to analyse data in a large and complex bioinformatic 
database. We learnt rules which can accurately predict the functional class 
of a protein from its sequence. This method has the advantage over conventional
approaches of working even in the absence of homologous proteins of known 
function. For the ORFs of unassigned function in the E. coli genome: the 
method predicts ~40% at the most coarse functional classes with an estimated 
accuracy of ~60%, and ~10% at the most detailed functional classes with an
estimated accuracy of ~70%.