Codex @codex  0

Two desirable properties are similar primer melting temperatures, allowing one annealing condition to support both primers efficiently, and specific binding with little self-complementarity or mutual complementarity, especially at the $3^{\prime }$ ends, to reduce off-target products and primer dimers. Suitable length and a balanced GC content help achieve these goals. A long homopolymer, strong hairpin or repeated genomic binding site would compromise a candidate even if its calculated melting temperature looked acceptable.

Base composition changes both the initial buoyant density and the size of the BUdR-induced density increment. The fraction-number axis runs toward lower density. The unlabelled viral peak lies at a lower fraction number than the corresponding host peak, consistent with viral DNA having a higher GC content. BUdR replaces thymidine in newly synthesized strands, so relatively AT-rich DNA incorporates more brominated nucleoside per unit length and undergoes a larger density increase.

Consequently the more AT-rich host DNA can overtake the viral DNA in the fully substituted class: the host fully heavy peak is slightly farther toward the dense end, although host unlabelled DNA was less dense. The hybrid peaks are intermediate. A useful qualitative model is ${\rho }_{LL}={\rho }_0(g)$, ${\rho }_{LH}={\rho }_0(g)+\Delta \rho (g)/2$ and ${\rho }_{HH}={\rho }_0(g)+\Delta \rho (g)$, where $g$ is GC fraction, ${\rho }_0$ increases with $g$, and $\Delta \rho$ decreases as the thymidine fraction decreases. Unequal substitution efficiencies could contribute, but different replication mechanisms are unnecessary. These are equilibrium density positions, not molecular-length measurements.