Please use Han et al (2016) Oncotarget 11269 to cite OASIS 2.

Benchmark

We provide the benchmark of statistical tests newly implemented in OASIS 2.OASIS 2 results are the same as the results obtained by SurvCurv, JMP, or STATA.

Samples

Samples of lifespan dataset used to benchmark can be accessed from link below. All raw data are provided as tab-delimited format of OASIS 2, JMP and STATA data format.

SampleDescriptionDownloadsReference
#1 C. elegans Wild-type vs. fat-6 OE OASIS 2  JMP  STATA Han et al., 2016
#2C. elegans Wild-type vs. daf-16 RNAi OASIS 2  JMP  STATA Gharbi et al., 2013
#3M. musculus Female vs. Male OASIS 2  JMP  STATA Strong et al., 2008

Comparison of Maximal Lifespan

  • Statistical significance (P-values)

      SampleFisher's Exact TestBoschloo's TestMann-Whitney U Test
      OASIS 2 SurvCurv JMP STATA OASIS 2 JMP STATA OASIS 2 JMP STATA
      #1 0.0008 0.0008 0.0008 0.001 0.029 . . 0.0009 0.0009 0.0009
      #2 3.9E-14 2.27E-09 < 0.0001* 0.000 7.6E-08 . . 1.2E-14 < 0.001* 0.0000
      #3 0.36 0.36 0.36 0.231 0.41 . . 0.329 0.329 0.328

  • 2x2 Contingency Table

      To compare the fraction or lifespan value of longevity outliers, each population in each condition is categorized into one of two subgroups: either lifespan above the estimated percentile or lifespan below the estimated percentile. 90 percentile is used to divide sub-group.

      SampleOASIS 2SurvCurv
      #1
      Lifespan above 90th percentile
      YesNo
      Wild-type15101
      fat-6 OE2117
      Lifespan above 90th percentile
      YesNo
      Wild-type15101
      fat-6 OE2117
      #2
      Lifespan above 90th percentile
      YesNo
      Wild-type34115
      daf-16 RNAi0239
      Lifespan above 90th percentile
      YesNo
      Wild-type34115
      daf-16 RNAi0239
      #3
      Lifespan above 90th percentile
      YesNo
      Female1182
      Male10116
      Lifespan above 90th percentile
      YesNo
      Female1281
      Male10116


  • Lifespan data of longevity outliers

      Lifespan data of longevity outliers (Z) is required to calculate Mann-Whitney U test. Following this method, we regarded lifespan as 0 if the the invidual's lifespan is not above 90th percentile.

      SampleLifespan data of longevity outliers
      #1
      #2
      #3

    		•

    Weighted Log-Rank Tests

  • Statistical significance (P-values)

      Sample Wilcoxon-Breslow-Gehan Test Tarone-Ware Test Peto-Peto-Prentice Test Fleming-Harrington Test
      OASIS 2 STATA OASIS 2 STATA OASIS 2 STATA OASIS 2 STATA
      #1 0.3127
      (χ2=1.02)      		 0.3127
      (χ2=1.02)      		 0.1855
      (χ2=1.75)      		 0.1855
      (χ2=1.75)      		 0.3327
      (χ2=0.94)      		 0.3327
      (χ2=0.94)      		 0.3038
      (χ2=1.06)      		 0.3038
      (χ2=1.06)
      #2 0
      (χ2=63.1)      		 0.0000
      (χ2=63.10)      		 0
      (χ2=72.11)      		 0.0000
      (χ2=72.11)      		 0
      (χ2=58.97)      		 0.0000
      (χ2=58.97)      		 0
      (χ2=63.07)      		 0.0000
      (χ2=63.07)
      #3 1.8E-5
      (χ2=18.35)      		 0.0000
      (χ2=18.35)      		 4.9E-5
      (χ2=16.47)      		 0.0000
      (χ2=16.47)      		 1.9E-5
      (χ2=18.27)      		 0.0000
      (χ2=18.27)      		 1.9E-5
      (χ2=18.33)      		 0.0000
      (χ2=18.33)
    • Structural Bioinformatics Lab.