Unlike conventional (hard) computing which relies on precise 
mathematical and analytical methods, software computing can be 
characterized by computational approaches that mimic human mind
to reason and learn in an imperfect environment. Fuzzy sets, 
artificial neural networks, evolutionary algorithms, support 
vector machines, wavelets, simulated annealing, tabu search are
just a few examples of the components of soft computing. From the 
definition, it can be seen that soft computing is naturally 
suitable for solving the complex biological and medical problems.
The talk will be on the soft bioinformatics techniques developed 
in my laboratory for the past few years on various subjects of 
biological sequence analysis. The ultimate goal is to develop 
bioinformatics methods that allows discovery of biological 
insights. This goal was pursued by combining soft computing 
with amino acid property-based approaches in the hope that 
bioinformatics algorithms are not only working but also 
accountable. Topics discussed in the talk include protein motif 
discovery, allergenic protein prediction, protein subcellular 
localization, protein family classification, missense amino acid 
mutation and the classical multiple sequence alignment problem. 
We will describe applications of wavelet analysis, support vector
machines and tabu search to the above problems.