Abstract 

In recent years, new efficient experimental techniques, especially in
the area of DNA sequencing, have led to a tremendous growth in available
biological data. Many large sequence databases are already publicly available
on the internet, and information is added at a spectacular rate. The extensive
human genome project is only one of the many sources of this information. 
It is
widely recognized that the mere gathering of data is not sufficient and that its
biological interpretation is of the utmost importance. Unfortunately, the
development of methods for in-terpreting the data is not keeping up with the
tempo with which the data is accumulated.  
It is clear that many types of
questions can only be asked by a computational analysis, and computer science
has become an integral part of the research involving biological sequences (of
DNA, RNA, or proteins). The research area combining biology and computer
science is known as bioinformatics. Conventional computer methods and
algorithms have been applied quite successfully in this area, but the often
enormous amounts of data to be analyzed and the complexity of biological systems
leave many interesting problems beyond the reach of conventional approaches.
The challenging computational problems of bioinformatics provide interesting
opportunities for applying methods from the field of artificial intelligence.
In this paper, the emphasis is on discussing how methods from the field of
uncertainty in AI can be relevant for some challenging problems of
bioinformatics.  Some necessary background information on molecular genetics in
general and the human genome project in particular is provided at the beginning
of the paper.