PCR in Action

we are ready to apply PCR to our simulated data set. For each training set absorbance matrix, A1 and A2, we will find all of the possible eigenvectors. Then, we will decide how many to keep as our basis set. Next, we will construct calibrations by using ILS in the new coordinate system defined by the basis set. Finally, we will use the calibrations to predict the concentrations for our validation sets. 

All we need do to calculate all the possible eigenvectors and eigenvalues is feed the data into an appropriate software package. So, we will begin the discussion with the question of how many of the eigenvectors to keep. 

We have an advantage in this situation. We know that our data contain 4 components, plus a small linear baseline, and some nonlinearities. This would lead us to expect that our data would show a rank of 5 or 6. Four dimensions are required to span the variations from the 4 different spectral components, and an additional 1 or two would be needed to span the variations due to the baselines and non-linearities. It gets a bit tricky when we realize that the baseline and non-linearity effects are rather small. Statistical indicators such as the 2-way F-test on REV s are not always able to destinguish factors that span small, but 

Let s see what the 2-way F-test tells us. Table 9 contains the REV s and F ratios for the data in the two training sets A1 and A2. The details of calculating the F ratios, and determining the values for the numerator and denominator are discussed in Appendix D. 

At the 10% level, these modified F ratios indicate that we need 5 factors for either Al or A2. At the 5% level, we get an indication of 4 factors for Al and 3 factors for A2. 

---

Principal Component Regression PCR in Action Partial Least-Squares PLS in Action The Beginning... 

The theoretical and practical features of PCR and the specific problems involved in its possible application to the pharmaceutical market lead to the following recommendable courses of action ... 

In addition to bcl-2, another hitherto completely unexpected target for the actions of chronic lithium and VPA has been identified from the mRNA RT-PCR DD study described above. Another clone, also derived from a transcript whose levels were increased by both lithium and VPA, shows very strong homology to a human mRNA-binding protein, the AUH protein ([54, 55] Genbank accession number X79888). BESTFIT analysis revealed 83.2% sequence homology between this rodent clone and the human AUH protein [54—56]. 

Enzymes from these organisms find numerous applications beyond PCR. Certain ones can modify plant fibers or break down proteins or fats. Heat-stable enzymes with these digestive properties are very attractive to the food processing industry. Properly controlled, their action can make prepared foods more palatable, and they can be employed where ordinary enzymes fail. In food processing, all operations must be carried out under sterile conditions. Frequently, the easiest way to maintain sterility is to keep the... 

---