I have done Comet test for a different exposures with xenobiotics to bivalves. I think that the time from Dna damage started depends on different elements, for example, the concentration of contaminant, the type of contaminant or where you detected the DNA damage (tissues, organs).
Actually, as you know, it's not in my major but I guess you may find something in this article. They measured DNA damage and DNA repair using Saccharomyces cerevisiae as model system:
Quantitative DNA Damage and Repair Measurement with the Yeast Comet Assay
that depends on the system you use for assessing the damage, as virtually all factors will vary in between different cell lines. We have observed an (small) instant response, an increase over the course of approx. 3 h and a return to negative control potential within approx. 12 hours after UV irradiation of cultured cells.
Something else you should consider especially for benzo[a]pyrene is bioactivation. If you use a cell line without significant biotransformation capacity, you will not observe any damage at all, as the genotoxic metabolites are not formed. Directly acting compounds such as alkylating agents are generally preferable in such cell lines.
A few years ago, we planned to exposed fish to this genotoxicant model B[a]P complementary to our experiment (MMS exposure) but finally we didn't. I remember we designed a protocol with sampling after 6h of exposure / 2 days / 7 days at 3 concentrations. The first one was an environmentally relevant one, a high concentration and an intermediate one (according to the litterature about long-term in-vivo exposure). It would be interesting to include a depuration stage in your experiment (at least a week and 2 sampling events). You should have a look on our paper published last year in Ecotox & Env safety, we designed a protocol to characterize our biomarker. You might read Wu et al. 2003 too which is very informative (see our references).
Let me know through researchgate or by email, I would be interested by your work.
Patrick is right on the money, as primary metabolic activation of B(a)P to active mutagenic product by endoplasmic membrane-bound CYP1s is the obligatory step determining any following DNA damage. B(a)P itself is totally inactive as a mutagen. To your very point, we'd conducted a long-term treatment of genetically different mice with B(a)P. First two week were found to be cricial in terms of B(a)P metabolic activation. Please find an abstract beneath, as well as a citation on our article published in Biochem Biophys Acta. Title: "A comparison of the activities of aryl hydrocarbon monooxygenase in liver microsomes from mice of different strains during prolonged 3,4-benzo(a)pyrene administration". By Ilya B. Tsyrlov, Vyacheslav V. Lyakhovich
ABSTRACT: The content and activity of the components of liver microsomal aryl hydrocarbon monooxygenase system change biphasically during long-term 3,4-benzo(a)pyrene administration of C57BL/6 mice as well as to (C57BL/6 X DBA/2)F1 hybrids. The first activity peak (4--14 days) is associated with the induction of aryl hydrocarbon monooxygenase by 3,4-benzo(a)pyrene; the second peak (70--84 days) is related to noninductive mechanism. In DBA/2 mice, the second peak is absent while the slight increase in aryl hydrocarbon monooxygenase activity observed on days 14--28 indicates the aberrant inductive capacity of 3,4-benzo(a)pyrene under its prolonged administration. It is suggested that the weak sensitivity to the blastogenesis caused by 3,4-benzo(a)pyrene observed in C57BL/6 mice and in (C57BL/6 X DBA/2)F1 hybrids is due to the high level of liver aryl hydrocarbon monooxygenase activity at the time of tumor appearance.