Registered report: A coding-independent function of gene and pseudogene mRNAs regulates tumour biology

The Reproducibility Project: Cancer Biology seeks to address growing concerns about reproducibility in scientific research by conducting replications of selected experiments from a number of high-profile papers in the field of cancer biology. The papers, which were published between 2010 and 2012, were selected on the basis of citations and Altmetric scores (Errington et al., 2014). This Registered report describes the proposed replication plan of key experiments from ‘A coding-independent function of gene and pseudogene mRNAs regulates tumour biology’ by Poliseno et al. (2010), published in Nature in 2010. The key experiments to be replicated are reported in Figures 1D, 2F-H, and 4A. In these experiments, Poliseno and colleagues report microRNAs miR-19b and miR-20a transcriptionally suppress both PTEN and PTENP1 in prostate cancer cells (Figure 1D; Poliseno et al., 2010). Decreased expression of PTEN and/or PTENP1 resulted in downregulated PTEN protein levels (Figure 2H), downregulation of both mRNAs (Figure 2G), and increased tumor cell proliferation (Figure 2F; Poliseno et al., 2010). Furthermore, overexpression of the PTEN 3′ UTR enhanced PTENP1 mRNA abundance limiting tumor cell proliferation, providing additional evidence for the co-regulation of PTEN and PTENP1 (Figure 4A; Poliseno et al., 2010). The Reproducibility Project: Cancer Biology is collaboration between the Center for Open Science and Science Exchange, and the results of the replications will be published in eLife. DOI: http://dx.doi.org/10.7554/eLife.08245.001


Introduction
The phosphatase and tensin homolog gene (PTEN) functions as a negative repressor of the PI3K/Akt survival pathway and is one of the most frequently deleted tumor suppressor genes in human cancer (Stambolic et al., 1998;Song et al., 2012). As a regulator of PI3K signaling, loss of PTEN results in over-activation of Akt, leading to unchecked cell proliferation, reduced apoptosis, and elevated tumor angiogenesis (Stambolic et al., 1998;Carracedo et al., 2008). In prostate cancer, decreases in PTEN protein expression, either by allelic deletion or functional loss caused by mutation and/or epigenetic modification, can lead to invasive prostate carcinoma (Trotman et al., 2003;Phin et al., 2013). In preclinical systems, the genetic restoration of PTEN induces apoptosis in cancer cell lines and has a significant negative effect on tumor growth in multiple in vivo models (Li et al., 1998;Lu et al., 1999;Tian et al., 1999;Chen et al., 2011). In contrast, clinical efforts to restore PTEN functionality have instead focused on targeting kinases in the PI3K pathway, including PI3K, Akt, and the mammalian target of rapamycin (Hopkins and Parsons, 2014). However, the development of PI3K targeting drugs has been complicated by the limited tolerability of current pharmacological treatments as well as tumor heterogeneity (Gerlinger et al., 2012;Bauer et al., 2014).
It is increasingly apparent that a complex regulatory network exists between the diverse RNA species pervasive in the human transcriptome. MicroRNAs (miRNAs) are small non-coding RNAs that bind to *For correspondence: tim@ cos.io complementary sequences in the 3′ untranslated regions (UTR) of target messenger RNAs (mRNA), resulting in transcriptional downregulation of the target gene (Sen et al., 2014). Meng and colleagues showed that PTEN was repressed by miR-21, one of the most frequently upregulated miRNAs in cancer, in hepatocarcinoma cells, suggesting that the oncogenic potential of miR-21 occurs via the downregulation of PTEN expression (Chan et al., 2005;Meng et al., 2006;Volinia et al., 2006;Meng et al., 2007;Si et al., 2007). Several miRNAs that target PTEN have since been reported (Jackson et al., 2014;Wang et al., 2015). While miRNAs play a functional role in silencing target gene expression, it is proposed that miRNAs themselves are subject to regulation by competing endogenous RNA (ceRNA) species, including pseudogenes, long non-coding RNAs, and circular RNAs Cesana and Daley, 2013). In plants, for example, the non-protein coding gene IPS1 sequesters miRNAs away from their mRNA targets, thereby leading to an accumulation of target transcripts (Franco-Zorrilla et al., 2007). Poliseno and colleagues proposed that pseudogenes, which are non-coding genomic DNA sequences closely related to parental genes, can modulate parental gene expression by influencing the available levels of miRNAs within a cell (Poliseno et al., 2010;Cesana and Daley, 2013). However, the extent and manner that ceRNAs can exert a consequential effect on the repression of targets for that miRNA is currently unclear (Broderick and Zamore, 2014). Recently, Denzler and colleagues analyzed the stoichiometric relationship of miR-122 and target sites in adult mouse liver and reported that the natural abundance of target sites exceeded miRNAs, making the ceRNA hypothesis unlikely (Denzler et al., 2014).
PTENP1 is a pseudogene that shares close homology with PTEN, including the ability to bind miRNAs (Fujii et al., 1999). To determine whether PTEN and PTENP1 expression levels are modulated by miRNA activity, Poliseno and colleagues first established that the PTEN-targeting miRNAs miR-19b and miR-20a were able to target both PTEN and PTENP1 (Poliseno et al., 2010). As reported in Figure 1D, overexpression of miR-19b and miR-20a in prostate cancer cells resulted in a significant decrease in PTEN and PTENP1 mRNA transcription. This is supported by additional studies demonstrating that overexpression of either miR-19b or miR-20a in cancer cell lines resulted in reduced PTEN mRNA levels and protein expression (Luo et al., 2013;Tian et al., 2013;Wu et al., 2014). The ability of miR-19b and miR-20a to target PTEN in prostate cancer was further confirmed by Tay et al. (2011). These key findings established that PTEN and PTENP1 are regulated by interactions with miRNA in multiple cancer cell types and will be replicated in Protocol 1.
In Figure 2F-H, Poliseno and colleagues tested the phenotypic consequences of PTENP1 downregulation by specifically targeting PTEN and/or PTENP1 expression. Downregulation of PTENP1 in DU145 prostate cancer cells resulted in a significant decrease in both PTEN and PTENP1 mRNA levels and protein expression (Figure 2G-H; Poliseno et al., 2010). Furthermore, downregulation of PTENP1 profoundly accelerated the proliferation of DU145 cells ( Figure 2F), with silencing of both PTEN and PTENP1 having an additive effect (Poliseno et al., 2010). These experiments will be replicated in Protocols 2, 3, and 4. Recently, Tay and colleagues reported that PTEN-ceRNAs, including CNOT6L and VAPA, phenocopied PTENP1 activity, as downregulation of these non-coding transcripts in prostate and colon cancer cells were also able to modulate PTEN expression, Akt activity, and cell growth . Additionally, other PTEN-ceRNAs that regulate PTEN expression have been reported in brain, breast, and skin cancers (Lee et al., 2010;Karreth et al., 2011;Sumazin et al., 2011). Further to this, PTENP1 antisense RNA has been reported to regulate PTEN transcription and mRNA stability, suggesting a model where the PTENP1 pseudogene has biomodal functionality modulating PTEN (Johnsson et al., 2013).
As an extension of the findings reported in Figure 2 and further genomic analysis, Poliseno and colleagues demonstrated that the PTEN 3′ UTR regulates pseudogene expression, since overexpression of the PTEN 3′ UTR was found to de-repress PTENP1 expression and inhibited DU145 proliferation ( Figure 4A) (Poliseno et al., 2010). These experiments will be replicated in Protocols 5 and 6. These results were also confirmed by experiments by Yu and colleagues showing that overexpression of either PTEN or PTENP1 suppressed renal cancer cell proliferation (Yu et al., 2014). Further to this, the oncosuppressive properties of overexpressing PTENP1 3′ UTR have been reported in various cancer cells (Poliseno et al., 2010;Chen et al., 2015;Guo et al., 2015).  Figure 1D).

Confirmatory analysis plan
This replication attempt will perform the statistical analysis listed below.
■ Statistical Analysis: ○ Note: at the time of analysis, we will perform the Shapiro-Wilk test and generate a quantilequantile plot to assess the normality of the data. We will also perform Levene's test to assess homoscedasticity. If the data appear skewed we will perform the appropriate transformation in order to proceed with the proposed statistical analysis. If this is not possible we will perform the planned comparisons using the Wilcoxon-Mann Whitney test. ○ One-way MANOVA of normalized PTEN or PTENP1 mRNA fold change in siLuc,19b,or 20a siRNA transfected cells with the following planned comparisons using the ○ Compute the effect sizes of each comparison, compare them against the effect size in the original paper and use a random effects meta-analytic approach to combine the original and replication effects, which will be presented as a forest plot. ■ Additional exploratory analysis: ○ The same analysis described above will be performed with 36B4 normalized values, which serves as an independent normalization control not included in the original analysis.

Known differences from the original study
The PTEN and PTENP1 mRNA levels will be normalized with an independent control (36B4) in addition to ACTIN. All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
The cell line used in this experiment will undergo STR profiling to confirm its identity and will be sent for mycoplasma testing to ensure there is no contamination. Transfection efficiency will be recorded for each replicate and any transfection that does not contain >90% efficiency will be excluded and not continue through the rest of the procedure. If the efficiency in the first attempt(s) does not obtain >90%, then any modifications to the transfection protocol will be recorded and the procedure will be maintained for the remaining replicates. The sample purity (A 260/280 ratio) of the isolated RNA from each sample will be reported. The PTEN and PTENP1 mRNA levels will be normalized with an independent control (36B4). All the raw data, including the analysis files, will be uploaded to the project page on the Open Science Framework (OSF) (https://osf.io/yyqas) and made publically available.
Protocol 2: Cell growth assay following siRNA transfection This experiment tests the effect of siRNA mediated depletion of PTEN, PTENP1, or both on the growth of DU145 cells. It is a replication of Figure 2F.

Materials and reagents
Procedure Note c All cells will be sent for mycoplasma testing and STR profiling. c DU145 cells grown in complete RPMI 1640: RPMI 1640 supplemented with 2 mM glutamine, 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin at 37˚C and 6% CO 2 .
1. Seed 1.5 × 10 5 DU145 cells per well in a 12-well dish. Grow overnight. 2. Transfect with 100 nM siRNAs (siPTEN, siPTENP1, siPTEN Smartpool (siPTEN and PTENP1), or siLuc in separate wells) using Dharmafect 1 according to manufacturer's instructions or leave untransfected. Transfect control cells with siGLO RISC-free control siRNA according to manufacturer's instructions. Grow overnight. 3. Confirm that >90% of siGLO transfected control cells show fluorescence, indicating they were successfully transfected. a. If transfection is less than 90%, record efficiency for attempt, exclude attempt and do not continue with the rest of the procedure. Repeat procedure until >90% efficiency is obtained. b. If modification to transfection (step 2) is needed, record and maintain modified steps for remaining replicates. 4. The day after transfection, resuspend 2 × 10 5 siLuc, siPTEN, siPTENP1, siPTEN/PTENP1, or untransfected cells in 50 ml fresh media. Seed three wells of six sets of 12-well plates with 2 ml of each cell line. Each set of 12 well plates should have three wells of each cell line. Incubate overnight. 5. Fix one plate every 24 hr starting after overnight incubation (the first plate fixed will be called day 0). a. Wash wells once in PBS. Deliverables ■ Data to be collected:

Reagent
○ Images of fluorescence and phase/contrast of siGLO transfected cells. ○ Raw data of absorbance from plate reader. ○ Graph of relative cell number for each cell line over time. (Compare to Figure 2F).

Confirmatory analysis plan
This replication attempt will perform the following statistical analysis listed below.
■ Statistical Analysis: ○ Note: at the time of analysis, we will perform the Shapiro-Wilk test and generate a quantilequantile plot to assess the normality of the data. We will also perform Levene's test to assess homoscedasticity. If the data appear skewed we will perform the appropriate transformation in order to proceed with the proposed statistical analysis. If this is not possible we will perform the equivalent non-parametric test. ○ Two-way ANOVA comparing Day 5 absorbance in siLuc, siPTEN, siPTENP1, or siPTEN/PTENP1 transfected cells with the following planned comparisons using the ○ Compute the effect sizes of each comparison, compare them against the effect size in the original paper and use a random effects meta-analytic approach to combine the original and replication effects, which will be presented as a forest plot.
Known differences from the original study All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
The cell line used in this experiment will undergo STR profiling to confirm its identity and will be sent for mycoplasma testing to ensure there is no contamination. Transfection efficiency will be recorded for each replicate and any transfection that does not contain >90% efficiency will be excluded and not continue through the rest of the procedure. If the efficiency in the first attempt(s) does not obtain >90%, then any modifications to the transfection protocol will be recorded and the procedure will be maintained for the remaining replicates. All the raw data, including the analysis files, will be uploaded to the project page on the OSF (https://osf.io/yyqas) and made publically available. a. If transfection is less than 90%, record efficiency for attempt, exclude attempt and do not continue with the rest of the procedure. Repeat procedure until >90% efficiency is obtained. b. If modification to transfection (step 2) is needed, record and maintain modified steps for remaining replicates. 4. 24 hr after transfection, extract total RNA directly on the culture dish using TRI reagent and 1bromo-3-chloropropane according to manufacturer's instructions. 5. Treat RNA with DNAse following manufacturer's instructions. 6. Reverse transcribe 1 μg RNA/sample into cDNA using first-strand cDNA synthesis kit with primers following manufacturer's instructions.  Figure 2G, left). ○ Graph of fold change PTENP1 mRNA expression relative to siLuc. (Compare to Figure 2G, right).

Confirmatory analysis plan
This replication attempt will perform the following statistical analysis listed below.
■ Statistical Analysis: ○ Note: at the time of analysis, we will perform the Shapiro-Wilk test and generate a quantilequantile plot to assess the normality of the data. We will also perform Levene's test to assess homoscedasticity. If the data appear skewed we will perform the appropriate transformation in order to proceed with the proposed statistical analysis. If this is not possible we will perform the planned comparisons using the Wilcoxon-Mann Whitney test. ○ Compute the effect sizes of each comparison, compare them against the effect size in the original paper and use a random effects meta-analytic approach to combine the original and replication effects, which will be presented as a forest plot. ■ Additional exploratory analysis: ○ The same analysis described above will be performed with 36B4 normalized values, which serves as an independent normalization control not included in the original analysis.

Known differences from the original study
The PTEN and PTENP1 mRNA levels will be normalized with an independent control (36B4) in addition to ACTIN. All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
The cell line used in this experiment will undergo STR profiling to confirm their identity and will be sent for mycoplasma testing to ensure there is no contamination. Transfection efficiency will be recorded for each replicate and any transfection that does not contain >90% efficiency will be excluded and not continue through the rest of the procedure. If the efficiency in the first attempt(s) does not obtain >90%, then any modifications to the transfection protocol will be recorded and the procedure will be maintained for the remaining replicates. The sample purity (A 260/280 ratio) of the isolated RNA from each sample will be reported. The PTEN and PTENP1 mRNA levels will be normalized with an independent control (36B4). All the raw data, including the analysis files, will be uploaded to the project page on the OSF (https://osf.io/yyqas) and made publically available.

Protocol 4: Western blot of cells transfected with siRNA
This experiment utilizes western blot to assess the protein levels of PTEN after depletion of PTEN, PTENP1, or both. It is a replication of Figure 2H.
Sampling ■ Experiment to be repeated a total of five times for a minimum power of 80%. The original data are qualitative, thus to determine an appropriate number of replicates to initially perform, sample sizes based on a range of potential variance was determined. a. If transfection is less than 90%, record efficiency for attempt, exclude attempt and do not continue with the rest of the procedure. Repeat procedure until >90% efficiency is obtained. b. If modification to transfection (step 2) is needed, record and maintain modified steps for remaining replicates.

48 hr after transfection lyse cells transfected with siRNAs and uninfected cells in lysis buffer on ice
for 30 min. a. Lysis buffer: 50 mM Tris pH8.0, 1 mM EDTA, 1 mM MgCl 2 , 150 mM NaCl, 1% NP-40, 1 mM β-glycerophosphate, 1 mM Na 3 VO 4 , 1 mM NaF, protease inhibitors. 5. Gently sonicate protein lysate for 3 to 4 bursts for 5 to 10 s. Clear lysate by centrifugation at 10,000×g for 10 min at 4˚C. 6. Perform Bradford protein determination assay following manufacturer's instructions. 8. Transfer to nitrocellulose membrane (pre-wetted with methanol before use) at 25 V constant for 1-2 hr in 1× transfer buffer with 20% methanol following manufacturer's instructions. a. After transfer, stain membrane with Ponceau S solution following manufacturer's instructions to visualize transferred protein. Image membrane, then wash out the Ponceau stain (additional quality control step). 9. Perform western blotting with the following antibodies following manufacturer's instructions. Use 1× TBS for washes and blocking reagent recommended by manufacturer. a. rabbit anti-PTEN; use at 1:1000 dilution; 54 kDa. b. mouse anti-Hsp90; use at 1:1000 dilution; 90 kDa. 10. Detect signal with appropriate HRP conjugated secondary antibody followed by chemiluminescence following manufacturer's instructions. 11. Analyze scanned images using Image J software.
a. Equal-sized regions of interest (ROI) will be positioned on specific bands. b. Background will be located within each individual lane but not occupied by any other discrete band. c. Subtract background pixel intensity from ROI pixel intensity. d. Normalize PTEN values by Hsp90 values from the same sample. 12. Repeat independently four additional times.
Deliverables ■ Data to be collected: ○ Images of fluorescence and phase/contrast of siGLO transfected cells. ○ Images of Ponceau stained membranes and full films for all western blots with ladder. (Compare to Figure 2H). ○ Raw data file of ROI and background pixel intensities. ○ Normalize PTEN values for each sample.

Confirmatory analysis plan
This replication attempt will perform the following statistical analysis listed below.

■ Statistical Analysis:
○ Note: at the time of analysis, we will perform the Shapiro-Wilk test and generate a quantilequantile plot to assess the normality of the data. We will also perform Levene's test to assess homoscedasticity. If the data appear skewed we will perform the appropriate transformation in order to proceed with the proposed statistical analysis. If this is not possible we will perform the equivalent non-parametric test. ○ Two-way ANOVA of normalized PTEN levels in siLuc, siPTEN, siPTENP1, or siPTEN/PTENP1 siRNA transfected cells with the following planned comparisons using the Bonferroni correction: 1. siLuc compared to siPTEN. 2. siLuc compared to siPTENP1. 3. siLuc compared to siPTEN/PTENP1. 4. siPTEN/PTENP1 compared to siPTEN. 5. siPTEN/PTENP1 compared to siPTENP1. ■ Meta-analysis of effect sizes: ○ The replication data (mean and 95% confidence interval) will be plotted with the original reported data value plotted as a single point on the same plot for comparison.

Known differences from the original study
The original study used 12 well plates seeded with 1.5 × 10 5 DU145 cells per well, which was increased 2.5× to account for the difference in cell surface area. All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
The cell line used in this experiment will undergo STR profiling to confirm their identity and will be sent for mycoplasma testing to ensure there is no contamination. Transfection efficiency will be recorded for each replicate and any transfection that does not contain >90% efficiency will be excluded and not continue through the rest of the procedure. If the efficiency in the first attempt (s) does not obtain >90%, then any modifications to the transfection protocol will be recorded and the procedure will be maintained for the remaining replicates. Ponceau stained membranes will be used to assess completeness of transfer. All the raw data, including the analysis files, will be uploaded to the project page on the OSF (https://osf.io/yyqas) and made publically available.

Protocol 5: Quantitative PCR following PTEN 3′ UTR transfection
This experiment tests the effect of expressing the 3′ UTR of PTENP1 on mRNA expression levels of PTENP1. It is a replication of the left panel of Figure 4A.

Procedure
Note c All cells will be sent for mycoplasma testing and STR profiling. c DU145 cells grown in complete RPMI 1640: RPMI 1640 supplemented with 2 mM glutamine, 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin at 37˚C and 6% CO 2 .
Deliverables ■ Data to be collected: ○ Purity (A 260/280 ratio) and concentration of isolated total RNA from cells. ○ Raw data for all qPCR reactions. ○ Quantification of PTENP1 mRNA levels relative to ACTIN or 36B4. ○ Quantification of fold change PTENP1 mRNA levels relative to pCMV transfected cells. (Compare to Figure 4A, left panel).

Confirmatory analysis plan
This replication attempt will perform the following statistical analysis listed below.
■ Statistical Analysis: ○ Note: at the time of analysis, we will perform the Shapiro-Wilk test and generate a quantile-quantile plot to assess the normality of the data. We will also perform Levene's test to assess homoscedasticity.
If the data appear skewed we will perform the appropriate transformation in order to proceed with the proposed statistical analysis. If this is not possible we will perform the equivalent non-parametric test. ○ Unpaired two-tailed t-test of PTENP1 mRNA levels of pCMV transfected cells compared to pCMV/PTEN 3′ UTR transfected cells.
■ Meta-analysis of effect sizes: ○ Compute the effect sizes of each comparison, compare them against the effect size in the original paper and use a random effects meta-analytic approach to combine the original and replication effects, which will be presented as a forest plot. ■ Additional exploratory analysis: ○ The same analysis described above will be performed with 36B4 normalized values, which serves as an independent normalization control not included in the original analysis.

Known differences from the original study
The PTENP1 mRNA levels will be normalized with an independent control (36B4) in addition to ACTIN. All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
The cell line used in this experiment will undergo STR profiling to confirm their identity and will be sent for mycoplasma testing to ensure there is no contamination. The sample purity (A 260/280 ratio) of the isolated RNA from each sample will be reported. The PTENP1 mRNA levels will be normalized with an independent control (36B4). All the raw data, including the analysis files, will be uploaded to the project page on the OSF (https://osf.io/yyqas) and made publically available.

Protocol 6: Cell growth assay following PTEN 3′ UTR transfection
This experiment tests the effect of expressing the 3′ UTR of PTENP1 on cell growth. It is a replication of the right panel of Figure 4A.
Sampling ■ Experiment to be repeated a total of three times for a minimum power of 98%.

Confirmatory analysis plan
This replication attempt will perform the following statistical analysis listed below.
■ Statistical Analysis: ○ Note: at the time of analysis, we will perform the Shapiro-Wilk test and generate a quantilequantile plot to assess the normality of the data. We will also perform Levene's test to assess homoscedasticity. If the data appear skewed we will perform the appropriate transformation in order to proceed with the proposed statistical analysis. If this is not possible we will perform the equivalent non-parametric test. ○ Unpaired two-tailed t-test of Day 5 absorbance of pCMV transfected cells compared to pCMV/ PTEN 3′ UTR transfected cells. ○ Unpaired two-tailed t-test of AUC measurements (determined from day 0, 1, 2, 3, 4, and 5 for each replicate) of pCMV transfected cells compared to pCMV/PTEN 3′ UTR transfected cells. ■ Meta-analysis of effect sizes: ○ Compute the effect sizes of each comparison, compare them against the effect size in the original paper and use a random effects meta-analytic approach to combine the original and replication effects, which will be presented as a forest plot.
Known differences from the original study All known differences are listed in the materials and reagents section above with the originally used item listed in the comments section. All differences have the same capabilities as the original and are not expected to alter the experimental design.

Provisions for quality control
The cell line used in this experiment will undergo STR profiling to confirm their identity and will be sent for mycoplasma testing to ensure there is no contamination. All the raw data, including the analysis files, will be uploaded to the project page on the OSF (https://osf.io/yyqas) and made publically available.

Power calculations
For additional details on power calculations, please see analysis scripts and associated files on the OSF: https://osf.io/cd2yq/

Protocol 1
Summary of original data estimated from graph reported in Figure 1D:

Test family
■ Due to the large variance, these parametric tests are only used for comparison purposes. The sample size is based on the non-parametric tests listed above. ■ Two-way ANOVA: Fixed effects, special, main effects and interactions: alpha error = 0.05. ○ Due to a lack of raw original data, we are unable to perform power calculations using a MANOVA. We are using a two-way ANOVA to estimate sample size.

Test family
■ Due to the large variance, these parametric tests are only used for comparison purposes. The sample size is based on the non-parametric tests listed above. ■ 2 tailed t test, difference between two independent means, Bonferroni's correction: alpha error = 0.0125.

Protocol 2
Summary of original data estimated from graph reported in Figure 2F: ■ 2 tailed t test, difference between two independent means, Bonferroni's correction: alpha error = 0.01.

Day 5 values AUC values
Protocol 3 Summary of original data estimated from graph reported in Figure 2G: ○ Due to a lack of raw original data, we are unable to perform power calculations using a MANOVA. We are using a two-way ANOVA to estimate sample size.

Test family
■ Due to the large variance, these parametric tests are only used for comparison purposes. The sample size is based on the non-parametric tests listed above.

Protocol 4
Summary of original data reported in Figure 2H: The original data do not indicate the error associated with multiple biological replicates. To identify a suitable sample size, power calculations were performed using different levels of relative variance.

Test family
■ Two-way ANOVA: Fixed effects, special, main effects and interactions: alpha error = 0.05. 'Power calculations' performed with G*Power software, version 3.1.7 (Faul et al., 2007). ANOVA F test statistic and partial η 2 performed with R software, version 3.1.2 (R Development Core Team, 2014).

Test family
■ 2 tailed t test, difference between two independent means, Bonferroni's correction: alpha error = 0.01.
'Power calculations' performed with G*Power software, version 3.1.7 (Faul et al., 2007 Continued on next page