Meeting Abstract
Determining the biological functions of genes can greatly improve our understanding of evolutionary tendencies. By better understanding the relationships between mutations in DNA sequences and the occurrence or absence of particular traits or diseases, we can also greatly accelerate medical advancements – particularly in individualized medicine. However, determining these functional relationships requires complex and time consuming procedures, often including invasive tests in live animals. Here I introduce a method to infer candidate genes by analyzing large amounts of incomplete DNA sequences across large sets of species.
First, I construct gene trees for each considered locus. Loci need not be genes, but usually will be. As some data will not be available, the sets of species included in gene trees will vary. The consensus tree, however, will include all of them. I then compare each species’ position within the consensus tree against its position in each gene tree in which it is present and record its relative displacement. This results in an incomplete displacement matrix.
Simultaneously I derive a large number of morphology trees based on observations of morphological traits. These can be expressed as Boolean or numerical values. Similar to the previously described approach I obtain a morphological displacement matrix which may also be incomplete.
In a final step I explore existing correlations between the two matrices that indicate which genes might be directly or indirectly influencing particular traits. This method does not aim to determine gene functionality by itself, but it will facilitate investigative research by focusing efforts on likely candidate genes. It can be completed on limited computing architecture, even when analyzing very large sets of data.
