Biomedical Engineering Reference
In-Depth Information
with a phenotype of:
3 round : 1 wrinkled
Punnet's model was soon extended by others, including his associate G.H. Hardy, to a more
generalizeable form. For example, the Hardy-Weinberg Principle, proposed in 1908, states that, in
the absence of forces that change gene ratios in populations, when random mating is permitted, the
frequencies of each allele will tend to remain constant throughout the following generations. Work by
subsequent scientists, such as the British statistician Ronald Fisher, further quantified the
observations of Mendel, Punnet, Hardy, and others.
When Fisher applied his statistical methods to Mendel's work in the 1930s, he showed that Mendel's
figures were too perfect. With the small sample size used by Mendel, his findings, which agree with
the ratios predicted by Punnet's squares, would be unlikely to be observed. Whether this apparently
intentional error was the result of Mendel's manipulation of the data or, as some historians assert,
due to incorrect reporting by his support staff, is unknown.
In terms of complexity, Mendelian genetics, while a milestone in the development of our
understanding of genetics, pales in comparison to many of the statistical challenges of modern
molecular biology. Even though many researchers work with statistics through the special function
keys on their calculators or a dedicated statistical analysis program, the application of statistics is
much more than simple data analysis. For example, statistical methods provide the basis for modern
genomic and proteomic laboratory automation. Automating manual operations like pipetting not only
saves time but, properly implemented, automation can eliminate or minimize many sources of
variability and provide for a more robust experimental procedure. The rapid advances in
bioinformatics, such as sequencing most of the human genome, have been possible because of the
availability of statistical methods that compare and manipulate data representative of nucleotide
sequences and computer-enabled laboratory automation. Machines—perhaps more appropriately
referred to as robots—have been used to automate error-prone, manual procedures such as micro-
pipetting to the point that computer-based tools can quickly analyze the data they produce in the
time that it would have taken to simply set up a manual experiment.
Moving to the wet lab, sequencing machines generate data on thousands of base pairs per hour, and
microarray experiments can collect data on the expression of tens of thousands of genes in a few
