As evidenced by the considerable differences among species and clades, paracluster evolution is dramatic, with constant creation of new clusters, expansions of existing clusters, and dissolution of old ones, to the extent that many genes within paraclusters as well as whole paraclusters appear species specific at the level of analysis now possible. What is interesting in this regard is the common observation that it is the demands of cell-to-cell communication and environmental interactions, especially the presence of infectious agents and xenobiotics, which often appears to drive paracluster expansion exemplified by the expansion of ZNF clusters in defense against proliferation of retroviral LTRs. The genome of the microcrustacean, Daphnia pulex, is a prime example as the expression of paralogous genes, many of which are in tandem arrays, is associated with environmental perturbations. In turn, this raises the question of the extent to which formation of novel paraclusters contributes to the process of speciation itself by facilitating niche adaptations. Gene clusters were qualitatively identified as a product of chromosome walks by chaining together genes that are in proximity to one another and share a common structural annotation according to one of the five datasets tested. Each dataset was tested separately. The chaining algorithm was implemented in a forward direction along the chromosomes and each gene was evaluated for inclusion in no more than one chain. Although the genes in a chain can share multiple combinations of annotations, the chain itself is represented by the least common subset of Dasatinib 302962-49-8 annotations shared by all its members. A new gene is added to a chain only if it shares at least one of the common annotations; if it does not share all of the common annotations, its inclusion will reduce the least common subset of annotations accordingly. Because chains are constructed by moving in rank order along a chromosome, the combination of domains represented within a chain could be influenced by gene ordering. Knowing the TWS119 GSK-3 inhibitor frequency in each genome of all the possible annotation values in each of the five datasets, a p-value was obtained for each putative chain by applying the hypergeometric probability distribution. P-values were subsequently corrected for according to the number of possible chains given the sizes of both the chain and the relevant genome; this provided a random expectation value for a given chain based on the size of the genome in consideration. Two parameters of the chaining algorithm were tested empirically, the allowed size of a gap, and the total number of gaps allowed in a cluster. It was determined that limiting the total number of gaps was not critical to the total amount of clustering and was only done so to manage the computer runtime performance.