Comment on 'Quantum correlations are weaved by the spinors of the Euclidean primitives'.

I point out fundamental mathematical errors in the recent paper published in this journal 'Quantum correlations are weaved by the spinors of the Euclidean primitives' by Joy Christian.

A solution for the rare type match problem when using the DIP-STR marker system.

The rare type match problem is an evaluative challenging situation in which the analysis of a DNA profile reveals the presence of (at least) one allele which is not contained in the reference database. This situation is challenging because an estimate for the frequency of occurrence of the profile in a given population needs sophisticated evaluative procedures. The rare type match problem is very common when the DIP-STR marker system, which has proven itself very useful for dealing with unbalanced DNA mixtures, is used, essentially due to the limited size of the available database. The object-oriented Bayesian network proposed in Cereda et al. [7] to assess the value of the evidence for general scenarios, was not designed to deal with this particular situation. In this paper, the model is extended and partially modified to be able to calculate the full Bayesian likelihood ratio in presence of any (observed and not yet observed) allele of a given profile. The method is based on the approach developed in Cereda [5] for Y-STR data. Alternative solutions, such as the plug-in approximation and an empirical Bayesian methodology are also proposed and compared with the results obtained with the full Bayesian approach.

Separable measurement estimation of density matrices and its fidelity gap with collective protocols.

We show that there exists a gap between the performance of separable and collective measurements in the qubit mixed-state estimation that persists in the large sample limit. We characterize the gap with sharp asymptotic bounds on mean fidelity. We present an adaptive protocol that attains the separable measurement bound. This protocol uses von Neumann measurements and can be easily implemented with current technology.

No time loophole in Bell's theorem: the Hess-Philipp model is nonlocal.

Hess and Philipp recently claimed [(2001) Proc. Natl. Acad. Sci. USA 98, 14224-14227 and 14228-14233] that proofs of Bell's theorem have overlooked the possibility of time dependence in local hidden variables, hence the theorem has not been proven true. Moreover they present what is claimed to be a local realistic model of the EPR correlations. If this is true then Bell's theorem is not just unproven, but false. We refute both claims. First, we explain why time is not an issue in Bell's theorem, and second, we show that their hidden variables model violates Einstein separability. Hess and Philipp have overlooked the freedom of the experimenter to choose settings of a measurement apparatus at will: any setting could be in force during the same time period.

Excision repair of nitrogen mustard-DNA adducts in Saccharomyces cerevisiae.

The bifunctional alkylating anticancer drug nitrogen mustard forms a variety of DNA lesions, including monoadducts and intrastrand and interstrand crosslinks. Although it is known that nucleotide excision repair (NER) is important in processing these adducts, the role of the other principal excision repair pathway, base excision repair (BER) is less well defined. Using isogenic Saccharomyces cerevisiae strains disrupted for a variety of NER and BER genes we have examined the relative importance of the two pathways in the repair of nitrogen mustard adducts. As expected, NER defective cells (rad4 and rad14 strains) are extremely sensitive to the drug. One of the BER mutants, a 3-methyladenine glycosylase defective (mag1) strain also shows significant hypersensitivity. Using a rad4/mag1 double mutant it is shown that the two excision repair pathways are epistatic to each other for nitrogen mustard sensitivity. Furthermore, both rad14 and mag1 disruptants show elevated levels of nitrogen mustard-induced forward mutation. Measurements of repair rates of nitrogen mustard N-alkylpurine adducts in the highly transcribed RPB2 gene demonstrate defects in the processing of mono-adducts in rad4, rad14 and mag1 strains. However, there are differences in the kinetics of adduct removal in the NER mutants compared to the mag1 strain. In the mag1 strain significant repair occurs within 1 h with evidence of enhanced repair on the transcribed strand. Adducts however accumulate at later times in this strain. In contrast, in the NER mutants repair is only evident at times greater than 1 h. In a mag1/rad4 double mutant damage accumulates with no evidence of repair. Comparison of the rates of repair in this gene with those in a different genomic region indicate that the contributions of NER and BER to the repair of nitrogen mustard adducts may not be the same genome wide.

Predictors of hospital acquired heel pressure ulcers.

The purpose of this study was to evaluate predictors of hospital acquired heel pressure ulcers. A prospective cohort study of hospitalized patients was conducted (N = 291). Subjects were enrolled by one team and followed by another team that was blind to initial assessment information. Initial assessment included demographics, Braden scale, and other variables found in the first study to be statistically significant. Ongoing evaluation involved heel assessment only. Univariate analysis yielded 15 statistically significant variables. Using multivariate logistic regression, subject's with a potential problem on the Braden Friction and Shear item (p = 0.01) and who were more frequently moist on the Braden Moisture item (p = 0.007) were more likely to develop heel ulcers (chi-square 30.52, df 3, p = 0.00001). Receiver Operator Characteristic (ROC) curves were plotted for the Braden scale and multiple other scoring systems. ROC curves were virtually identical using all new scoring systems as compared to the original Braden scale. No new scoring system was identified that led to a clinically significant improvement in sensitivity/specificity over the total Braden scale. While not perfect, the Braden scale may currently be the best predictive tool for heel pressure ulcer development.

Analyzing bivariate continuous data grouped into categories defined by empirical quantiles of marginal distributions.

Epidemiologists sometimes study the association between two measurements of exposure on the same subjects by grouping the original bivariate continuous data into categories that are defined by the empirical quantiles of the two marginal distributions. Although such grouped data are presented in a two-way contingency table, the cell counts in this table do not have a multinomial distribution. We describe the joint distribution of counts in such a table by the term empirical bivariate quantile-partitioned (EBQP) distribution. Blomqvist (1950, Annals of Mathematical Statistics 21, 539-600) gave an asymptotic EBQP theory for bivariate data partitioned by the sample medians. We demonstrate that his asymptotic theory is not correct, however, except in special cases. We present a general asymptotic theory for tables of arbitrary dimensions and apply this theory to construct confidence intervals for the kappa statistic. We show by simulations that the confidence interval procedures we propose have near nominal coverage for sample sizes exceeding 60 for both 2 x 2 and 3 x 3 tables. These simulations also illustrate that the asymptotic theory of Blomqvist (1950) and the methods that Fleiss, Cohen, and Everitt (1969, Psychological Bulletin 72, 323-327) give for multinomial tables can yield subnominal coverage for kappa calculated from EBQP tables, although in some cases the coverage for these procedures is near nominal levels.

Non-response models for the analysis of non-monotone ignorable missing data.

We discuss a new class of ignorable non-monotone missing data models-the randomized monotone missingness (RMM) models. We argue that the RMM models represent the most general plausible physical mechanism for generating non-monotone ignorable data. We show that there exists ignorable missing data processes that are not RMM. We argue that it may therefore be inappropriate to analyse non-monotone missing data under the assumption that the missingness mechanism is ignorable, if a statistical test has rejected the hypothesis that the missing data process is RMM representable. We use RMM models to analyse data from a case-control study of the effects of radiation on breast cancer.

Increased resistance to N,N',N"-triethylenethiophosphoramide (thiotepa) in cells expressing the Escherichia coli formamidopyrimidine-DNA glycosylase.

Thiotepa (N,N',N'-triethylenethiophosphoramide) is an alkylating agent used in cancer chemotherapy. A reaction pathway by which thiotepa alkylates purified DNA involves hydrolysis to aziridine, which forms N7-aminoethyl guanine and aminoethyl adenine. These lesions are repaired in Escherichia coli by the formamidopyrimidine-DNA glycosylase (Fpg) protein. To determine whether such lesions are formed by thiotepa in mammalian cells, we have overexpressed the E. coli Fpg protein in Chinese hamster ovary cells. The transfected cells were more resistant to the lethal and mutagenic effects of thiotepa than the parental cells. The number of replication-blocking lesions formed by thiotepa, measured by quantitative PCR analysis, was lower in the transfected cells. These results show that expression of the Fpg protein increases the cell resistance to thiotepa and suggest that this compound produces ring-opened guanines, which are involved in its cytotoxic action.

Targeted A --> T and G --> T mutations induced by site-specific deoxyadenosine and deoxyguanosine adducts, respectively, from the (+)-anti-diol epoxide of dibenz[a,j]anthracene in M13mp7L2.

The studies described in this report directly examined the mutagenicity in Escherichia coli of both a deoxyadenosine (dAdo) and a deoxyguanosine (dGuo) adduct derived from (+)-anti-dibenz[a,j]-anthracene-3,4-diol 1,2-epoxide [(+)anti-DB[a,j]A-DE] that were site-specifically placed in a single-stranded M13mp7L2 replication vector. An 11-base oligonucleotide (5'-CTC ACG CTT CT-3') containing either a single (+)anti-DB[a,j]A-DE--trans-N2-dGuo or (+)anti-DB[a,j]A-DE--trans-N6dAdo adduct was successfully incorporated into single-stranded M13mp7L2 plasmid via ligation. In vitro studies using E. coli DNA polymerase I (Klenow fragment)indicated that both adducts were effective blocks for polymerase action. E. coli strains JM103 and JM103 uvrA6 were subsequently transformed with control (unadducted) and adduct-containing M13mp7L2 constructs followed by analysis of progeny DNA. In both JM103 and JM103 uvrA6 cells, plaque yields were markedly reduced with adduct containing vectors compared to control vectors. Activation of the inducible bacterial DNA repair system (SOS) by UV light only slightly increased the number of plaques recovered from either bacterial strain transformed with adduct-containing vectors. Targeted mutations were obtained with both adduct-containing vectors in both bacterial strains, whereas no mutations were detected in plaques recovered from control M13mp7L2 vectors. In JM103 cells, (+)anti-DB[a,j]A-DE--N6-dAdo induced exclusively A --> t transversions and (+)anti-DB[a,j]A-DE--N2-dGuo induced exclusively G --> T transversions. In JM103 uvrA6 cells, similar targeted transversion mutations were also obtained except that a few C deletions (i.e., aprroximately 10% of the mutations) were detected immediately 3' to the dAdo adduct. While mutagenesis was SOS dependent in JM103 cells [<0.15% (-SOS) vs approximately 1.3% (+SOS)], it appeared to be SOS independent in JM103 uvrA6 cells (approximately 1-2% in the presence or absence of SOS induction). It is argued that adduct-induced G --> T mutations can be rationalized by either misinformational or noninformational mechanisms. In contrast, A --> T mutations are unlikely to arise via a misinformational pathway, which provides the strongest support to date that bulky DNA adducts can induce mutations via a noninformational pathway.

Curve fitting of labeled platelets: gamma model expanded for severe pathology.

To obtain an adequate curve fitting of labeled platelets also in case of severe pathology the multiple hit model of Murphy & Francis can be used, however, with addition of fractions of n (argument of the gamma function) and expansion of the domain of n towards zero and negative values. The 'hazard rate function' provides the instantaneous death rate for a platelet that has survived up to a given time. Applied to the multiple hit function, three kinds of hazard rate will be found: for n > 1 the death rate increases with age (absent or moderate pathology); for n = 1 the death rate is independent of age (death at random); for n < 1 the death rate decreases with age (probably severe pathology).

Construction of Escherichia coli vectors containing deoxyadenosine and deoxyguanosine adducts from (+)-anti-dibenz[a,j]anthracene diol epoxide at a defined site.

Dibenz[a,j]anthracene (DB[a,j]A) is a carcinogenic polycylic aromatic hydrocarbon, which is metabolically activated through the formation of bay region diol epoxides. Site-specifically modified M13mp19-based vectors containing a single (+)-anti-dibenz[a,j] anthracene diol epoxide [(+)-anti-DB[a,j[A-DE]-deoxyguanosine (dGuo) or -deoxyadenosine (dAdo) adduct were constructed. Four-base oligonucleotides, 5'-HOTGCA-3' and 5'-HOCATG-3', corresponding to the central four base pairs in the PstI and SphI restriction endonuclease sites, respectively, in the multiple cloning region of M13mp19, were reacted in solution with (+/-)-anti-DB[a,j]A-DE. The resulting adducted oligonucleotides were separated and purified using reverse-phase HPLC. Several different singly adducted oligonucleotides were isolated, consisting of the various cis and trans addition products of the (+) and (-) enantiomers of the diol epoxide bound to dGuo or dAdo in the oligonucleotides. 5'-HOTGCA-3' containing the (+)-anti-DB[a,j]A-trans-N2-dGuo adduct [T(DB[a,j]A-N2)GCA] and 5'-HOCATG-3' containing the (+)-anti-DB[a,j]A-trans-N6-dAdo adduct [C(DB[a,j]A-N6)ATG) were selected for subsequent ligation into M13mp19 vectors that had been constructed with a corresponding four base gap in the minus strand. Both unmodified and adducted oligonucleotides were successfully ligated into the M13mp19 vectors, [yields: unmodified -TGCA-M13mp19 (approximately 32%) and -CATG- M13mp19 (approximately 42%); adducted T(DB[a,j]A-N2)GCA-M13mp19 (approximately 13%) and C(DB[a,j]A-N6)ATG-M13mp19 (approximately 12%)]. The dAdo adduct-containing vector was characterized. The presence of a dAdo-DNA adduct at the recognition site of SphI inhibited restriction by SphI.(ABSTRACT TRUNCATED AT 250 WORDS)

Further analysis of c-Ha-ras mutations in papillomas initiated by several polycyclic aromatic hydrocarbons and papillomas from uninitiated, promoter-treated skin in SENCAR mice.

In this study we analyzed the mutations in c-Ha-ras from skin papillomas initiated with benzo[a]pyrene (B[a]P), 7-methylbenz[a]anthracene (7-MBA), and 10-fluoro-7-methylbenz[a]anthracene (10-F-7-MBA) and from papillomas induced by treatment with tumor promoter alone. Among the papillomas induced by treatment with tumor promoter alone, 56% (nine of 16) had mutations in c-Ha-ras. These mutations were found primarily in codon 61 and included both A182-->T and A182-->G mutations. In addition, one promoter-induced tumor had a G35-->A mutation in codon 12, and one had a G37-->C mutation in codon 13. The other promoter-induced papillomas did not have detectable mutations in codons 12, 13, or 61. Most of the B[a]P-initiated papillomas (77%; 10 of 13) did not have detectable mutations in c-Ha-ras codons 12, 13, or 61. However, three of these B[a]P-initiated papillomas had c-Ha-ras codon 13 mutations; one had a G37-->C transversion and two had G38-->T transversions. Most of the 7-MBA-initiated tumors and all of the 10-F-7-MBA-initiated tumors had an activated c-Ha-ras gene [nine of 10 (90%) and 11 of 11 (100%), respectively]. These mutations were almost exclusively A182-->T transversions in codon 61 except for two 7-MBA-initiated papillomas that had G37-->C transversions in codon 13. The results suggest that more than one mechanism may contribute to activation of c-Ha-ras by polycyclic aromatic hydrocarbons (PAHs) in mouse skin. Furthermore, the absence of c-Ha-ras mutations in most B[a]P-initiated papillomas, as well as in a significant fraction of those induced by tumor promoter alone, suggests that there may be other molecular targets involved in tumor initiation by PAHs in mouse skin.

Mutagenic specificity of the (+)anti-diol epoxide of dibenz[a,j]anthracene in the supF gene of an Escherichia coli plasmid.

This study was designed to examine the mutagenic specificity of (+)anti-dibenz[a,j]anthracene 3,4-diol-1,2-epoxide ((+)anti-DB[a,j]A-DE) in SOS-induced repair-proficient Escherichia coli ES87 (delta pro-lac, strA)/F' (pro+, lac1Q, lac1am26, lacZ delta M15). The plasmid pUB3, which contains the mutation target gene, supF, was modified with (+)anti-DB[a,j]A-DE in vitro (two to five adducts/plasmid) and then transformed into bacteria by electroporation. The spontaneous mutation frequency for unmodified pUB3 in uninduced cells was about 2 x 10(-6) and for SOS-induced cells, about 8 x 10(-6). The spontaneous supF- mutations were primarily insertions, deletions, and frameshifts. The mutation frequency for (+)anti-DB[a,j]A-DE-modified pUB3 was about 8 x 10(-6) and about 32 x 10(-6) for uninduced cells and SOS-induced cells, respectively. (+)anti-DB[a,j]A-DE induced primarily point mutations in supF in SOS-induced cells. GC-->AT transitions were the major mutations observed in SOS-induced cells (37%). GC-->TA (21%) and GC-->CG (8.6%) transversion mutations were also observed, whereas mutations at AT base pairs were rare (1.9%). Furthermore, a large number of tandem GC/GC-->AT/AT transition mutations were also observed (about 15% of all mutations in SOS-induced cells). Taken together, single and tandem GC-->AT mutations accounted for slightly over half (about 51%) of the mutations observed in SOS-induced cells. These results demonstrated that (+)anti-DB[a,j]A-DE was mutagenic in repair-proficient E. coli; however, unlike other polycyclic aromatic hydrocarbons that induce primarily transversion mutations, (+)anti-DB[a,j]A-DE caused mostly GC-->AT transitions.

Evaluation of sustained hyperplasia and other short-term tests as predictors of tumorigenic potential in oil products.

Some oil products are known to cause skin tumors following long-term application while others do not. The ability to predict which ones might cause tumors is important. Development of reliable short term tests which can accurately predict tumorigenic potential of oil products is needed to avoid the high cost and long time required for traditional animal bioassays. Several short term tests were evaluated for their ability to predict tumorigenic potential of 10 coded oil samples and results were compared to results of mouse bioassays. Analytical determinations of PAC content (DMSO extraction) and 3-6 ring PAC content were also made for each of the 10 samples for further comparison. Tests which showed good correlation with bioassay results and thus were considered good predictors of tumorigenic potential were: Sustained Epidermal Hyperplasia as measured by epidermal thickness, Nuclear Area of epidermal basal cells, Modified Ames Test and DMSO extraction for PAC content. Tests which did not show good correlation with bioassay results and which were not considered good predictors of tumorigenic potential were: Polymorphonuclear leukocyte (PMN) infiltration into the dermis, Unscheduled DNA Synthesis in epidermal cells and Changes in Nuclear DNA Content of CHO cells.

Analysis of point mutations in murine c-Ha-ras of skin tumors initiated with dibenz[a,j]anthracene and derivatives.

This study was designed to evaluate the point mutations in the murine c-Ha-ras gene of skin papillomas induced by initiation with dibenz[a,j]anthracene (DB[a,j]A), its bay-region anti-diol epoxide ((+/-)anti-DB[a,j]A-DE), and a 7,14-dimethyl analogue (7,14-diMeDB[a,j]A). Recent studies (Nair RV, et al., Chem Res Toxicol 4:115-122, 1991) in our laboratory have revealed both deoxyguanosine (dGuo) and deoxyadenosine (dAdo) adducts formed from the anti- and syn-diol epoxides of DB[a,j]A in cultured mouse epidermal cells after exposure to this hydrocarbon. Using PCR amplification and direct sequencing, we found specific A182----T transversion mutations (eight of 10 tumors) in codon 61 of c-Ha-ras in papillomas induced by initiation with DB[a,j]A. Analysis of papillomas generated by initiation with the more biologically potent analogue 7,14-diMeDB[a,j]A revealed that five of five tumors exhibited A182----T transversions in codon 61. The nature of the changes in the two DB[a,j]A tumors not showing codon 61 mutations in Ha-ras is currently not known since these tumor DNAs also did not possess c-Ha-ras mutations at codons 12, 13, or 59. Interestingly, papillomas produced by initiation with (+/-)anti-DB[a,j]A-DE also possessed A182----T transversion mutations in codon 61 of c-Ha-ras (five of five tumors). These data suggest that dAdo adducts derived from both parent hydrocarbons may play an important role in their tumor-initiating activity and possibly implicate a specific diol epoxide-dAdo adduct in this process.

Multistate life-tables and regression models.

"A survey is given of the use of modern statistical techniques in event history analysis, and in particular in the study of multi-state life-tables in demography. Emphasis is placed on the interplay between partial likelihood and nonparametric maximum likelihood based methods, a) when analysing semi-Markov models or models with repeated spells, and b) in frailty models for unobservable heterogeneity."

Characterization of covalently modified deoxyribonucleosides formed from dibenz[a,j]anthracene in primary cultures of mouse keratinocytes.

Identification of various deoxyribonucleoside adducts formed in primary cultures of mouse keratinocytes exposed to dibenz[a,j]anthracene (DB[a,j]A) is presented. A preliminary analysis of the DNA adducts formed from 7-methyldibenz[a,j]anthracene (7MeDB[a,j]A) also is presented. Cultures of keratinocytes obtained from dorsal skins of female SENCAR mice were exposed to 0.5 microgram of tritium-labeled hydrocarbons/mL of medium for 24 h. The total DNA binding was 2.23 +/- 0.54 and 5.28 +/- 0.97 pmol of hydrocarbon/mg of DNA for DB[a,j]A and 7MeDB[a,j]A, respectively. These binding values represented the radioactivity associated with the modified deoxyribonucleosides separated from the normal deoxyribonucleosides on Sephadex LH-20 columns following enzymatic digestion of isolated DNA. Treatment of keratinocytes with DB[a,j]A produced adduct peaks corresponding to marker adducts derived from trans addition of both deoxyguanosine as well as deoxyadenosine residues to the (+) enantiomer of the anti-diol epoxide where the deoxyadenosine adducts were predominant. In addition, DNA adduct peaks corresponding to markers of trans and cis addition, respectively, of deoxyguanosine and deoxyadenosine to the (+)-syn-diol epoxide were also noted in these chromatograms. A major DNA adduct in cells exposed to DB[a,j]A was tentatively identified as resulting from the addition of deoxyadenosine to DB[a,j]A-5,6-oxide. Several other later eluting DNA adduct peaks, not corresponding to any of the marker adducts, were also present in these chromatograms. In comparison, when cells were exposed to the more biologically potent 7-methyl analogue, at least 12 DNA adduct peaks were consistently observed in HPLC chromatograms.(ABSTRACT TRUNCATED AT 250 WORDS)

Relationship between DNA adduct formation and unscheduled DNA synthesis (UDS) in cultured mouse epidermal keratinocytes.

Primary cultures of mouse epidermal keratinocytes from SENCAR mice were treated with 7,12-dimethylbenz(a)anthracene (DMBA), benzo(a)pyrene [B(a)P], (+/-)7 beta-8 alpha-dihydroxy-9 alpha, 10 alpha-epoxy-7,8,9,10-tetrahydrobenzo(a)pyrene [(+/-)anti-BPDE], and (+/-)7 beta, 8 alpha-dihydroxy-9 beta, 10 beta-epoxy-7,8,9,10-tetrahydrobenzo(a)pyrene [(+/-)syn-BPDE] to examine the relationship between DNA adduct formation and the induction of unscheduled DNA synthesis (UDS). DNA adducts were measured as pmol hydrocarbon bound per mg of DNA, and UDS was quantitated autoradiographically as net grains per nucleus. A good correlation was observed between the levels of UDS detected and the amount of DNA adducts present in the cell population when comparing similar compounds within the linear dose-response range of 0.005 micrograms/ml-0.25 micrograms/ml. A higher rate of UDS for a given level of DNA adducts was interpreted as an increased efficiency of DNA repair. In some cases, an increase in the efficiency of DNA excision repair correlated with lower tumor-initiating activity. For this family of PAH, the concentration below which UDS could no longer be detected was approximately 0.01 microgram/ml. However, DNA adducts were measurable at concentrations of 0.01 and 0.005 micrograms/ml. The limits of detection of the current UDS assay in the SENCAR MEK culture system occurred at hydrocarbon adduct levels of approximately 10 pmol/mg DNA, or approximately 1 adduct per 3 x 10(5) bases. Additionally, the UDS assay was unable to detect DNA repair induced by the weakly carcinogenic PAHs, dibenz(aj)anthracene and 7-methyl-dibenz(aj)anthracene. The UDS assay did detect DNA repair by the more strongly carcinogenic PAH, 6-methylcholanthrene. These results suggest that the present UDS assay with MEKs is a useful assay for the rapid screening of potential genotoxic agents. However, the limits of sensitivity are such that the current assay may be unable to detect a low level of DNA damage induced by some weakly genotoxic (carcinogenic) agents. In addition, while the limits of sensitivity determined in these experiments apply to the polycyclic aromatic hydrocarbon class, other classes of genotoxic compounds such as alkylating agents or crosslinking agents may exhibit different thresholds of detection.

Distribution of covalent DNA adducts in mouse epidermal subpopulations after topical application of benzo(a)pyrene and 7,12-dimethylbenz(a)anthracene.

The distribution of benzo(a)pyrene [B(a)P] and 7,12-dimethylbenz(a)anthracene (DMBA):DNA adducts was examined in five different subpopulations of SENCAR mouse epidermal cells separated based on buoyant density in continuous gradients of 61.5% Percoll. Three fractions consisted of primarily basal cells (Fractions 3 to 5), while two less dense fractions (Fractions 1 and 2) consisted of primarily differentiating keratinocytes. The levels of B(a)P and DMBA:DNA adducts were examined at 1 h, 6 h, 24 h, 72 h (except DMBA), and 28 days after a single topical application of an initiating dose. Among the basal cell subpopulations, the level of covalent B(a)P:DNA adducts in Fraction 5 cells was significantly higher (P less than 0.05) than Fractions 3 and 4 at every time point examined. On the other hand, B(a)P:DNA adduct levels in Fraction 5 were only significantly higher than Fraction 2 at 6 h and 72 h and not significantly different from Fraction 1 at any time point. With DMBA, no significant differences were initially observed in the levels of covalent DNA adducts among the various Percoll fractions at 1 h and 6 h after treatment. However, at 24 h and at 28 days. Fraction 5 cells had significantly higher (P less than 0.05) levels of covalent DMBA:DNA adducts than Fractions 1 to 4. To explore whether the observed differences in DNA adduct levels were due to differences in metabolic activation, we examined the levels of covalent adducts among epidermal subpopulations after topical application of (+/-)-anti-benzo(a)pyrene-7,8-diol-9,10-epoxide (anti-BPDE). Interestingly, 3 h after treatment with anti-BPDE, significantly higher (P less than 0.05) levels of binding were found in Fraction 5 compared with Fractions 1 to 4. High-pressure liquid chromatographic analyses of B(a)P and DMBA:DNA adducts 6 h and 24 h after treatment did not show any significant differences in adduct profiles among the various subpopulations. These results demonstrate the presence and persistence of hydrocarbon:DNA adducts in all epidermal subpopulations isolated on continuous Percoll gradients for at least 28 days after treatment. Furthermore, of the three basal cell subpopulations, the most dense cells (Fraction 5) developed the highest DNA adduct levels within 24 h and retained these higher levels over 28 days. Finally, differences in DNA adduct levels among epidermal subpopulations do not appear to result from different metabolic capabilities of the cells. The potential significance of these results is discussed in terms of the process of skin tumor initiation.