Table 1. Numb re‐expression affects the tumorigenicity and CSC content of Numb tumors, but not of Numb+ tumors
TumorTreatmentNr. of cells injectedFreq. SCs/CICs95% CIP‐value
1051045 × 103103102101
Outgrowths/injections
Numb tumorsT3Ctr2/24/43/42/80/80/41:3,2901:1,550–1:6,9850.013
T3Numb2/23/62/60/20/20/21:13,9991:5,803–1:33,772
T4Ctr2/24/43/41/80/80/41:3,9541:1,857–1:8,4200.017
T4Numb2/24/82/80/20/20/21:15,7511:7,110–1:34,892
Numb+ tumorsTCCtr2/22/25/70/80/20/21:5,0901:2,396–1:10,8110.949
TCNumb2/22/23/51/60/80/41:4,8971:2,116–1:11,335
TDCtr2/22/24/61/80/20/21:4,6601:2,147–1:10,1150.649
TDNumb2/22/23/61/80/20/21:6,0961:2,685–1:13,839
  • Numb or Numb+ MECs, reconstituted with Numb‐DsRed (Numb) or mock‐infected (Ctr), were orthotopically transplanted at limiting dilutions in NGS mice. The number of outgrowths per number of injected pads (Outgrowths/injections) is indicated. The complete statistical analysis of the experiment, with the frequency of CICs and with 95% confidence intervals (CI), is shown. The frequencies were calculated by Poisson statistics, using the “StatMod” software package for the R computing environment (http://www.R-project.org), as previously described (Shackleton et al, 2006), and a complementary log‐log generalized linear model (two‐sided 95% Wald confidence intervals or, in case of zero outgrowths, one‐sided 95% Clopper–Pearson intervals). The single‐hit assumption was tested as recommended and was not rejected for any dilution series (> 0.05).