Today it is clear that individuals are identifiable through their unique biological profiles, and evidence of this - for example, from gene expression or microbiome datasets - continues to accumulate.
This method was illustrated in (Tibshirani et al., 2005) with examples taken from gene expression datasets and mass spectrometry.
We cannot, however, exclude the possibility that unmeasured or residual confounding, for example from genes or aspects of social background, may contribute to explain our findings.
Representative examples derived from gene-set enrichment analysis (GSEA) are shown in Figure 3.
Examples from other gene variants exist, for example the GNB3 825C>T polymorphism [39], where the largest association is seen among carriers of heterozygous genotypes.
Locus-specific examples from targeted gene studies also demonstrate a role for DNA methylation changes in AD and include observed disease-associated differences at ribosomal RNA gene promoters [ 38], genome-wide LINE-1 elements [ 39], and known AD susceptibility genes [ 40].
Why seemingly infinite number of state space (for example, genotypes in gene regulation networks) can result in a finite of number of functional states (for example, phenotypes from gene regulations)?
For example, findings from gene mutant mice show that p16INK4a and p19ARF, two senescence-associated proteins that contribute to stem cell failure during aging of the hematopoietic, neural and cardiac systems [ 7- 9], do not play a comparable role in restraining oogenesis in adult females [ 10].
To build a robust classifier, an important next step in our PUDI algorithm is to further extract the likely positive examples LP and the likely negative examples LN from genes in the U-RN which are near the positive and negative classification boundary.
