Global analysis of biochemical and biophysical data.

Publisher Summary A major transformation of the amount of biological information can be effected, which can be obtained from a set of experiments by direct incorporation of scientific constraints into the analysis of the data. The chapter has provided a "cookbook" approach to develop (and using) nonlinear least-squares analysis programs. The term “scientific constraints” is meant to denote direct application (into the analysis program) of the mathematical constraints imposed on the data through experimental design and accessory information. When first approaching a data analysis problem, what general class of computer algorithms should be applied to the data needs to be determined. The relationship between experimental observables and biological information in almost all of biochemistry and biophysics is nonlinear. Therefore, the best methodology to fit a nonlinear model to experimental data needs to be determined. The role of the classic biochemical data transformations should be exclusively for graphical representation of the data. All of the data analysis should be performed in a nonlinear fashion on the actual (raw) data. The nonlinear data analysis requirements of different laboratories in the field of biochemistry and biophysics vary greatly. For the laboratory, which occasionally needs to fit small numbers of data (say 10–500 data points) to a model, data can be easily described analytically (for example, sums of exponentials, simple binding isotherms); prepackaged nonlinear data analysis programs will probably be sufficient.