Upload StatisticalAnalysisPrompt.txt with huggingface_hub

StatisticalAnalysisPrompt.txt

@@ -0,0 +1,44 @@
+
+
+Statistical analysis prompt v02 10_31_24 JM
+
+
+You are a reviewer of a paper submitted to a high-quality scientific journal, and you want to apply the highest standards in assessing the submission. You are particularly concerned about the paper presenting results and data in a way most directly related to its central hypothesis. 
+
+From this point of view, please systematically assess the statistical and data presentation aspects of the paper. Please identify the main findings of the paper, and for each of these in turn, where relevant, address the following points one by one. 
+
+Please regard the paper as presenting evidence for and against a specific mechanism by which the E4 allele of the APOE gene increases risk of Alzheimer’s disease compared with the common E3 allele. The central hypothesis is that in humans, the E4 allele of APOE affects aspects of the blood brain barrier, relative to E3, with specific downstream consequences in the brain directly linked to increase disease risk. 
+
+In doing this, please distinguish between data and analysis improvements it would be nice to have and those aspects which are really key to the correctness or otherwise of the central hypothesis, and so affect your confidence in the main conclusions. That is, assign a relevance to every criticism of the methodology and results that you address, on a scale of 0 - 5. 
+
+Please be specific and detailed in your analysis. For example, if effect sizes are missing for some result, please state exactly which ones. 
+
+1. Statistical significance: are appropriate and clear statistical tests used for each aspect of the results? Are appropriate P values and other confidence measures provided? What are they? What further tests, if any, should be included? 
+
+2. Are relevant effect sizes clearly given, with uncertainties.
+
+3. Where needed, are appropriate multi-test corrections made? (this is a common deficiency of biology papers).
+
+4. Are appropriate controls used throughout? Are appropriate comparisons with the E4 reference state included? 
+
+5. Are the data presented in an optimal way with respect to the hypothesis or hypotheses? Are separate plots used for quantities that should be directly compared? For example, two quantities are each compared with a third and not with each other, when the latter comparison is most relevant.
+
+6. Are there any missing desirable data? 
+
+7. Are consistent axis scalings used for plots that should be comparable? 
+
+8. Are data presenting with controls normalized to 1.0, so potentially masking control behavior inconsistency across experiments? 
+
+9. Biological significance: results may be statistically significance, but still unlikely to be biologically significant. For example, small expression differences for highly expressed genes may be detected with very high confidence, but be too small to be biologically relevant. Please assess each finding in these terms.  
+
+10. Is any procedure, such as a Grubb test, used to remove outlier data points, and if so, is such removal clearly justified, rather than perhaps being used to improve the apparent statistical significance of the results?
+
+11. Are inappropriate assumptions of variable independence used?
+
+12. Are there any other potential concerns about the statistics and presentation of the data?
+
+13. Given the above, what confidence (on a scale of 0 to 5) would you assign to each of the principal findings and the main hypothesis or hypotheses?  
+
+Please summarize the main strengths and weaknesses emerging from the analysis.
+
+Use pymupdf4llm to analyze the paper