1 ~.5 experimentally deduced internal constraints.'' Furthermore, their use is restricted to highly specialized teams with a high level of physicochemical and mathematical training and who have at their disposal large computer facilities. However, for most initial studies the use of traditional and less sophisticated methods for the prediction of the secondary structure or the polar/apolar character along the protein sequence, 4'6-9 is highly informative. The application of these methods is usually simple, either being used manually or by computer-assisted approaches. Here we shall examine the reliability and accuracy of several simple predictive methods by applying them to the conformational analysis of a very well-known protein, chicken lysozyme. The limitations and the precautions that must be taken in the application of these methods will be stressed. Moreover, several simple computer programmes for the application of those predictive methods and for the semi-quantitative analysis of structural homologies among proteins will be discussed. A practical demonstration of these abilities will be made by comparing lysozyme and a-lactalbumin, two proteins considered as 'text-book' examples of homology. We present here a review of those methods which may also be used as a practical class for students of biochemistry and molecular biology since it requires discussion and work on different basic aspects of protein chemistry. Thus, the relationships between sequence-conformation and function is stressed. Moreover, important concepts such as protein differentiation, conservative substitutions, differential evolution of protein segments at the sequence and conformational level, and intronexon location and function, amongst others, are also discussed. The work was carried out in collaboration with a student (MF) in the last year of her degree in biochemistry. Background From the 1970s many empirical methods for the prediction of conformational traits of a protein from its amino acid sequence were developed (see ref 6-9 for reviews). These methods can be classified into two groups: those based on probabilistic studies and those based on a physicochemical approach. Most of these methods start from the same initial observation: each of the twenty usual amino acid residues in a protein has a particular potential to participate in different secondary structures (eg ahelix, 6-strand, reverse turn, random coil) or in buried or exposed stretches of it. Quantitative scales were deduced from statistical analysis of the structures of a large., mngr°up of proteins of known three-dimensional conformanon ' (probabilistic method) or from the quantification, using different techniques, of the polar/apolar character of each residue and its tendency to become exposed or buried lz (physicochemical method). To start the analysis, using probabilistic methods, an analytical window including a small number of residues (three to seven) is usually selected. The average (or positional, in certain cases) probability of the participation of this group of residues in a given secondary structure is calculated, and a probability profile for this conformational trait is elaborated by moving the window along the sequence of the protein.l°'13-15The regions of the protein which present an averaged probability above a neutral (or threshold) value for a particular trait are predicted to possess this trait. When one region has positive probabilities for more than one conformation trait, qualitative 1° or mathematical discriminant rules s'1°'16 are given to select the most probable. The physicochemical-based methods follow a similar practical analytical procedure, with a moving window to elaborate the 'hydropathy profile' and a given threshold to distinguish the polar/apolar characters for a given residue or protein stretch. 17-19 This profile is directly related to the overall (secondary and tertiary) folding of the protein. Some of these physicochemical methods also try to predict definite secondary structures through the observation of repeated polar and nonpolar patterns of residues (eg 1-2-5 or 1-4-5 for a-helix and alternate polar/non-polar for 13-strand) along the sequence. 12,~)-22