See post for proposed agenda, how I did from last week’s goals, and the new goals for this upcoming week.

Did I do What I said I would?

Yes.

See post where I listed my goals: July 8, 2025

Aaaaand… yes I did do what I intended to do! See today’s earlier post where I link out all the files (note: you will only be able to access the GitHub repository and files if you are a collaborator, as it is currently set to private). Link to post: July 18, 2025.

Agenda for today:

  1. Go over the annotations I did
  2. Come up with next steps for results to share at WDA
  3. Come up with target journal to submit Aspen’s paper that’s currently on biorxiv: Uncovering Host-Parasite Dynamics: Gene Expression Shifts in Hematodinium-infected Chionoecetes bairdi in Response to Temperature Change

Meeting Notes:

Will be placed here after the meeting occurs at 12-12:30pm today

Question: Why does each count matrix have some instances where the gene counts are “0”s across all 6 libraries? Need to figure that out.

  • Look at just BPs
  • Or go to GO Slim and pick categories to compare
  • Pull out every row that has immune in it “grep”
    Create smaller tables of just “immune” processes

Next step:

  • Just pull out rows that have “immune” in them
    Calculation on counts
    Give sum of all 6 columns (stars)
    Another column that does mean
    Another column that does std error
    Could pick 4 genes (house-keeping - or one gene)
    What people use (NADPH or ribosomal gene)
    Could normalize based on that
  • Following step:
    Find common genes
    Find them based on shared names
    Swissprot IDs
  • Could in a separate effort:
    Do a join on that for all three species
    Join based on swissprot ID
    How many are shared

To Do:

Multispecies

  1. For each count matrix of each species:
    • pull out all genes that are annotated with “immune”
    • get sum of counts across all 6 libraries
    • get mean of counts
    • get std error
  2. Compare across all three species:
    • join each annotated count matrix by the uniprot accession ID column and see what is shared
    • can also see what is unique to each species
    • can then filter down smaller to just “immune” processes
  3. Get GO Slim terms for each species annotated count matrix.

Crab Paper:

Find some journal options to think about.