False positives

  

Figure 6: Expression profiles of four genes consistently mis-classified with respect to the MYGD class of cytoplasmic ribosomal proteins. Each figure shows the expression profile for a single gene, along with standard deviation bars for the class of cytoplasmic ribosomal proteins. Ticks along the X-axis represent the beginnings of experimental series, as described in Figure 1. The genes in Figures (a)-(c) are false positives; i.e., the SVM places them in the class but MYGD does not. The gene in Figure (d) is a false negative.

Many of the false positives in Table 6 are known from biochemical studies to be important for the same functional class assigned by the SVM, even though MYGD has not included these genes in their functional class. For example, YAL003W and YPL037C, two false positives assigned repeatedly to the cytoplasmic ribosome class, are not strictly ribosomal proteins; however, both are important for proper functioning of the ribosome. YAL003W encodes a translation elongation factor, EFB1, known to be required for the proper functioning of the ribosome [Kinzy and J. L. Woolford, 1995]. YPL037C, EGD1, is part of the nascent polypeptide-associated complex, which has been shown to associate with the ribosome and help target nascent polypeptides to several locations including the endoplasmic reticulum and mitochondria [George et al., 1998]. The cell ensures that expression of these proteins keeps pace with the expression of ribosomal proteins, as shown in Figure 6(a) and (c). Thus, the SVM classifies YAL003W and YPL037C with ribosomal proteins.

The respiration complexes class provides another example, YML120C, NADH:ubiquinone oxidoreductase. In yeast, YML120C replaces respiration complex 1 [Marres et al., 1991], and while it does not pump protons across the mitochondrial inner membrane, this gene is crucial to the proper functioning of the respiration complexes. Without YML120C, the respiration chain is unable to transfer high energy electrons from NADH to ubiquinone, and respiration stops [Marres et al., 1991,Kitajima-Ihara and Yagi, 1998]. In the proteasome class YGR048W, ufd1, is classified by the SVM as part of the proteasome class. While YGR048W is not strictly part of the proteasome, it is necessary for proper functioning of the ubiquitin degradation pathway [Johnson et al., 1995], which delivers proteins to the proteasome for proteolysis.

Other examples include the classification of members of the tricarboxylic-acid (TCA) pathway, YPL262W and YKL085W, as members of the respiration chain complexes. While MYGD separates the tricarboxylic-acid (TCA) pathway and the respiration chain complexes, these two classes are known to be tightly coupled in the production of ATP. This relationship is represented in the expression data by the similarity between the expression profiles of the two classes, as shown in Figure 5 and leads the SVM to classify YPL262W and YKL085W as respiration complexes. Thus, while MYGD considers these two classes separate, both the expression data and other experimental work suggest that YPL262W and YKL085W have important roles to play in the function of the respiration complexes.