In jestarling/starUtils: Tools for Automating Project and Report Assembly

library(knitr)
opts_chunk$set(fig.path="Figures/projDesc_")

Project Overview

Personnel

This project is for PI NAME in the Department of DEPT NAME. Bioinformatics personnel are NAMES.

Motivating Question

What are we trying to understand?
How does this relate to basic biology and/or patient care?
How will the data being collected supply information relevant to addressing these issues?

Timeline Constraints

Are there deadlines we're trying to meet? When?

Protocol and/or Funding Source

Where are the resources coming from?

Project Folder Name and Location

This project is located in full path to this folder location is this on a shared drive? * is this a persistent repository?

Data and Methods

Raw Data Location

Where is the raw data? Supply a web link or relative path.
Is this a persistent repository?
When did we first get it?

Where is the relevant metadata (e.g., sample labels)? "As above" works, if this is true.

Where is the relevant clinical data?

Are we still waiting for any data? If so, when do we expect it?

Brief Data Description

What type of data is this? (E.g., RNA-Seq, methylation array)
What format is the data in? (E.g., Excel file, BAM files)
How many samples do we (should we?) have data for?

How do we expect this data to relate to the motivating question?

Experimental Design

What do we know about the experimental design?

What sample groups are to be contrasted?
Are there reasons we might need to be particularly concerned about batch effects and/or confounding?

Was the sample size chosen based on some type of power calculation? If so, what was the primary endpoint?

Analysis Outline

Data Gathering

What types of data gathering will be required?

Data Preprocessing

What types of data preprocessing (including data cleaning) will be required?
Can we specify the workflow, including software and metadata to be used?

Statistical Tests

What types of summary statistics will be provided?

What tests will we use for our comparisons? Two-sample t-tests? ANOVA? Cox PH models?

Do we need to worry about multiple testing? If so, do we have a rough idea of how we will address this? (E.g., fitting beta-uniform mixture models to the distribution of p-values, and selecting all variates significant at a specified false discovery rate (FDR))

Are there some types of outcomes which can be described as "success" or "failure" a priori?

Software

What software do we expect to use for the analyses described above? What versions?

Sanity Checks

Are there some differences we know should be there before we run the experiment? What are they? What direction should the differences be in?

Do we know what "no structure" should look like?

Do we expect a principal components analysis (PCA) plot to show clear differences other than those associated with the contrast of interest?

Do we need to consider training and test sets?

Do we have enough data to allow for determining an empirical null distribution based on permutation?

Figures

Are there specific figures we anticipate producing?

Results

What output files, reports, figures, and other deliverables do we expect to produce?

Do we have a rough timeframe / order for when these might be expected, at least for the first few?

Appendix

getwd()
sessionInfo()

jestarling/starUtils documentation built on May 30, 2019, 8:04 a.m.