Significance analysis of microarray (SAM) ‘outlier’ plots and heat maps from microarray analyses. (A) Four Caenorhabditis elegans full-genome microarrays were probed with cDNAs prepared from age-1(mg44)-F2 adults, mixed pairwise with cDNAs from adults bearing the age-1(hx546) allele (four independent cohorts of each strain), and four more arrays were probed with cDNAs from age-1(m333)-F2 adults, mixed pairwise with age-1(hx546) cDNAs (four cohorts of each strain). Data were combined from eight arrays (treating the two very long-lived strains as equivalent) and analyzed using SAM (Tusher et al., 2001). The SAM plot shows outliers in green if transcript levels are lower in the strong age-1 alleles (mg44 and m333) than in hx546, and in red if higher for the strong alleles, at a FDR of < 5%. (B) Three microarrays were probed with cDNAs from age-1(mg44)-F2 adults, mixed pairwise with cDNAs from N2DRM adults bearing the wild-type age-1 allele (three independent cohorts per strain), and three further arrays were probed with cDNAs from age-1(m333)-F2 adults, mixed pairwise with N2DRM cDNAs (three cohorts per strain). Data were combined from six arrays (treating the two very long-lived strains as equivalent) and analyzed by SAM as above. (C) Partial heat maps represent data from eight arrays probed with cDNA from age-1(hx546) adults (left), and four each probed with cDNAs from strong age-1 alleles mg44 and m333 (center). Expression ratios are shown for each array (right) as the log₂(mg44/hx546) or (m333/hx546); red intensity indicates positive log-ratios (higher expression in mg44 or m333), whereas green implies negative log-ratios (lower expression in mg44 or m333). Full heat maps are presented in Fig. S1. (D) Partial heat maps represent data from six arrays probed with cDNA from N2DRM adults (left), and three each probed with cDNAs from age-1 alleles mg44 and m333 (center). Expression ratios were calculated for each array (right) as the log₂(mg44/N2) or (m333/N2); red intensity indicates positive log-ratios (higher expression in mg44 or m333), whereas green implies negative log-ratios (lower expression in mg44 or m333). Full heat maps are presented in Fig. S2 (Supporting information).

The age-1(mg44)-F2 transcript profile differs from those of F1 adults or N2 dauer larvae. We retested 13 age-1-regulated genes, in an expanded strain panel. Strains are as described in the legend to Table 3, with the addition of age-1(mg44) F1 (fertile first-generation homozygotes), daf-2(e1370), a long-lived mutant of the insulin/IGF-1 receptor (G–M), and daf-16(mu86); age-1(mg44), in which a daf-16 deletion allele, mu86, replaces the m26 allele used in Tables 3 and 4. Double mutants with daf-16(mu86) were assessed only for those genes that had appeared to be incompletely (or negligibly) reverted by daf-16(m26) (F, H–M; indicated by black striping on green bars). Orange bars [at the right of (G–M)] represent N2DRM dauer larvae, induced by crowding on starved plates and confirmed by > 95% resistance to 1% sodium dodecyl sulfate for samples of 500 worms. Sample numbers (n = the number of independent biological preparations) were as follows: 8–11 groups of postgravid N2DRM adults, 4–5 each of age-1(hx546) and daf-16(m26); age-1(hx546), 3 of age-1(mg44) F1, 10–14 of age-1(mg44)-F2 adults, 5–6 each of daf-16(m26); age-1(mg44) and daf-16(mu86); age-1(mg44), 3–4 of N2DRM dauers, and 3–4 of daf-2(e1370).

Transcriptional changes in age-1(mg44) F2 adults are relatively stable with respect to age, and cannot be explained as a result of their physiological youthfulness. Transcript levels were assayed by real-time reverse-transcription polymerase chain reaction (RT-qPCR) as described in the legends to Table 3 and Fig. 2. Within each histogram, mean ± SEM are shown on a log₂ scale for steady-state transcript level, comparing wild-type N2DRM adults at three ages to age-1(mg44) F2 homozygotes at two adult ages. All transcript values were first normalized to β-actin mRNA (to adjust for variation in RNA inputs), and then each biological group was normalized to the mean value for N2DRM day-6 adults (to correct for run-to-run variation in C_t values). Ages are measured from the L4/adult molt; a table (inset) indicates the percent of adult lifespan represented by each age. To monitor longitudinal changes, two biological preparations of each group were assessed (fewer than used for Fig. 2 or Table 3). Significance of differences was ascertained by two-tailed Behrens-Fisher t-tests, appropriate to samples of unequal or unknown variance. The transcript levels observed in this experiment differ from the means seen in previous experiments (e.g. Tables 3 and 4), but those differences are not significant.

RNAi knock-down of genes downregulated in age-1(mg44) confers peroxide resistance and can extend life. Caenorhabditis elegans bearing (A) age-1(hx546) or (B) the wild-type age-1 allele were fed bacteria expressing the indicated dsRNA. RNAi treatment in (A), (B) and (D) commenced during early larval development of the parental generation, so that the tested generation underwent knockdown throughout life. Worms in (C) were fed dsRNA-expressing bacteria continuously from the L4/adult molt onward. Survival statistics are summarized in Tables 5 and 6.