Site Overview #

This is Anne Marie’s thesis work presented in an interactive and searchable website that was built from scratch. This website was developed to streamline the access to the entirety of my analysis, organized in a hierarchical fashion. The adaptable design allows for seamless additions, updates, and refinements. Detailed webpages feature multimodal content that support initial presentation and ongoing reference.

Key Features:

1. Tabbed Navigation for Figures: Facilitates comparative analysis across multiple data modalities and groups by efficiently organizing figures within tabs, minimizing screen clutter.
2. Interactive Plots: Dynamic plots with embedded details allow for real-time data exploration and insight generation.
3. Interactive Tables: Detailed data summaries are designed for immediate and tailored inquiries with the capability to filter, sort, and download.

Disclaimer: Please note the presence of some typos across the website, a result of preparing the text within Rmarkdown documents, which lack helpful features of standard word processors.

Technical Overview: This website was built entirely using R and Rmarkdown, which makes it easy to keep all information up to date. I utilized the blogdown library to convert Rmarkdown documents into markdown format. The Hugo framework generates a static site from these markdown files. I opted for a minimalistic theme that I customized to my preferences. Changes are pushed to GitHub, triggering AWS Amplify to detect these updates and automatically rebuild the site.

Cardiorespiratory Fitness Training Study #

Analysis Progress (non-exhaustive list):

• Working On
  • FST Analysis
  • Differential analysis interpretation of individual histone marks, ATACSeq, RNASeq, & chromatin states (pathway enrichment analysis, integration, etc.)
  • Untangle potential interaction between exercise and line
• To Do
  • Overlay FST sites with peaks and differential TF binding sites to understand their potential role in regulating gene expression
  • ATACSeq footprinting analysis (I already carried out a version of this using TOBIAS, but this will need to be revisited based on the Parker lab’s protocols to consider alternate alleles)
  • Preparation of F2 genotyping data and RNASeq as point of validation for findings from HCR/LCR FST analysis
• Completed
  • Data Processing
    • Alignment of RNASeq, ATACSeq, and CUTTag data to rn7 genome
    • Creation of count matrices (created consensus peak sets for CUTTag & ATAC datasets)
    • Assessed sample swaps in CUTTag data. Fortunately, sample ID in question was labeled accurately
    • Created blacklist region file custom to this data in this analysis - redid upstream analysis (call peaks, ataqv, create master peak regions and count matrices, peak annotations) for CUTTag & ATACSeq to consider this filter
    • Confirmed that exercise groups are correct in the metadata file used in the CUTTag snakemake pipeline processing directory (8/17/23)
    • Misc QC:
  • Differential Analysis
    • Latent variable discovery for data adjustment
    • Upset plots:
      • Overlap between significant line genes
  • Chromatin state discovery (ChromHMM)
• Benched
  • CUTTag QC exercise - annotate peaks called for individual samples to validate histone marks
    • Nothing well into the analysis has suggested that this is a possible issue

Misc #

Helpful resources I used to create this website #

• git lfs for tracking large png files: documentation
• This thread explains how to load the extended version of hugo on amplify, in the case a custom docker image is not used
• I used the custom docker image, weitzela/aws-amplify-tools:latest, in aws amplify to build the website. The Dockerfile I used to create the image loaded on aws amplify is located in this repository at resources/aws-amplify-docker/Dockerfile
  • Briefly, if the Dockerfile is created and works without error, you can simply run docker build -t username/image-name:version . && docker run username/image-name:version "bash" as a one liner to prepare it for use
  • Resources referred to to learn how to configure git lfs on amplify, which is required to load/show pictures tracked by git lfs on the url hosted site:
    • Dockerfile posts I referred to: one, two-a, two-b
    • Explanation of how to load git lfs from amazon-linux-extras: thread
    • Random: apparently git lfs has been added to amazon linux:2023, but at the time aws amplify only supported up to aws linux:2. This thread-unsuccessful explained how to install it during the build
  • Docker Troubleshoot: Iteratively create a docker file by…
    1. Building the image with commands you know work, in the directory the Dockerfile is located in: docker build -t image-name .
    2. Get the docker image ID: docker images
    3. Run an interactive instance of the image to test additional commands: docker run --rm -ti image-id. Hint: run whoami to test if you’re in the docker image.
    4. Add successful commands to the Dockerfile
    5. Repeat
  • Upload Image to Hub ( directions):
    1. Tag the image to include your docker username to allow permission to upload to repository: docker tag image-name:version username/image-name:version
       • Note: for this time I created the docker image, I ran docker run username/image-name:version "bash" before pushing to the hub
    2. Push to hub: docker push username/image-name:version