single hypothesis testing - Translational Medicine：假设检验单-转化医学.pptVIP

* Single Test Hypothesis Step 1: Hypothesis: Null hypothesis (no difference) Step 2: Experiment ? Biological Read-outs Step 3: Single Test (t test) ? P-values Step 4: Conclusion Validate/Reject Hypothesis Start-over with a different protein data set Step 1: Null Hypothesis Ho: Ho: Xh = Xd or d = 0 (No difference between the biological read-outs of healthy subjects and patients with a disease.) Patient (w/ disease) Healthy Patient (w/o disease) Protein … … Step 2: The experiment: Experiment yields biological read-outs of specific proteins for each patient. Patient 1 Patient 2 Patient 3 Patient 4 Protein 1 A A A A Protein 2 B B C C Protein 3 … … … … Protein 4 … … … … Protein 5 … … … … … … … … … Protein 100,000 … … … … Patients w/ disease Healthy Patients Biological read-outs (concentration of proteins) Step 3: Single test / t test: evaluates the difference between two means (healthy vs. sick) ? P-value Patient 1 Patient 2 Patient 3 Patient 4 Protein 1 A A A A Protein 2 B B C C Protein 3 … … … … Protein 4 … … … … Protein 5 … … … … … … … … … Protein 100,000 … … … … Patients w/ disease Healthy Patients Step 3: (continued) Patient 1 Patient 2 Patient 3 Patient 4 P-value Protein 1 A A A A Protein 2 B B C C … … … … … Protein 100,000 … … … … Patients w/ disease Healthy Patients Protein 1: averages: Patients 1 and 2 ≈ A Patients 3 and 4 ≈ A Difference = 0 Difference = B – C t-value ? P-value A t-value ? P-value B Protein 2: averages: Patients 1 and 2 ≈ B Patients 3 and 4 ≈ C P-value A P-value B *