Last Updated on 2024-10-17

Prot papers bot

This is the bot for @protpapers.bsky.social that was inspired and developed from bots created by https://github.com/roblanf/phypapers and https://github.com/JBGruber/r-bloggers-bluesky.

The bot runs twice a day to post new papers from PubMed and bioRxiv for subjects that are of interest to me, namely HLA immunopeptidomics, neoantigens and HDX-MS.

It’s written in R and uses Github Actions and YML file with cron to run workflow.

Here’s what I did if you want to make your own. There’s probably ways to make it better.

1. Set-up Bluesky

Follow these instructions from the phypapers readme:

“Obviously you need an account to post to. This part gets you set up on Bluesky, whether you have an existing personal account or not.

If you don’t have a Bluesky account: Go to https://bsky.app/, and click ‘Sign Up’ If you do have a Bluesky account, log in then go to Settings and click Add Account Leave the hosting provider as Bluesky Social Fill out your details to set up your new account: Protip: If you already use your gmail address for your account, you can just append to it to create a new account. E.g. if your personal account is porcelain.crab@gmail.com, you could use porcelain.crab+phypapers@gmail.com (the ‘+’ stays). This helps keep mail separate. Decide on a handle. Following flypapers’ lead, I suggest a short prefix followed immediately by papers, e.g. flypapers, phypapers, etc. This means we all know it when we see a literature bot. Click on your new profile, go to Edit Profile Username: I suggest making this prefix_papers e.g. fly_papers or phy_papers. As above, this helps everyone know what’s a literature bot Description: pretty obvious, but it’s always nice to know the human who runs it, so good to put your name there if you want to. It would be great if you could also put a link to these instructions on your literature bot - that way anyone who sees yours can also make their own. On my profile I just wrote: “Make your own literature bot with these instructions: https://github.com/roblanf/phypapers.”

There’s another step to set-up and get the app password and token, but we do that with the atrrr package: https://jbgruber.github.io/atrrr/index.html later on.

RSS feeds

For the RSS feeds I played around with the searches on PubMed and bioRxiv until I got something reasonable.

PubMed

I followed the phypapers readme again:

1. Go here: http://www.ncbi.nlm.nih.gov/pubmed/

2. Type in your favourite search terms remembering that wildcards are useful (e.g. phylogen* will match anything starting with phylogen, and logical operators can be really good, e.g. you can have phylogen* OR raxml OR splitstree.

3. Click search

4. Click the Create RSS link just below the search box

5. Name is something sensible

6. Set Number of items to be displayed to 100 7. Click Create RSS 8. Record the RSS URL somewhere

The somewhere bit for me is a character vector in R in bot.R.

After a bit of trial and error I ended up using search terms with flags for the title and abstract:

immunopeptidom*[tiab], hdx-ms[tiab], immunopeptidom*[tiab] AND neoantigen*[tiab].

bioRxiv

The full list of bioRxiv RSS feeds can be found on biorxiv alertsrss but you can see the ones I used in bot.R as a character vector.

# Vector of feeds of possible interest from bioRxiv, yields the last 30 days
brv_feeds <- c("http://connect.biorxiv.org/biorxiv_xml.php?subject=biochemistry",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=bioinformatics",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=biophysics",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=cancer_biology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=cell_biology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=developmental_biology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=genetics",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=genomics",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=immunology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=microbiology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=molecular_biology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=neuroscience",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=pharmacology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=physiology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=synthetic_biology",
               "http://connect.biorxiv.org/biorxiv_xml.php?subject=systems_biology")

Unlike PubMed the feeds provide all the papers the last 30 days and then have to filter them for your interests. I did this using stringr and a regex in bot.R like so:

brv_filt <- brv |> 
  filter(str_detect(item_title, "[Ii]mmunopep*|[Pp]eptidomi*|[Pp]eptidome|HDX-MS|([Pp]roteogenomics & [Nn]eoantigen)") |
         str_detect(item_description, "[Ii]mmunopep*|[Pp]eptidomi*|[Pp]eptidome|HDX-MS|([Pp]roteogenomics & [Nn]eoantigen)")) |> 
  mutate(link = str_extract(item_link,"^.*?[^?]*"))

Adapt the R script

The R script is based on the instructions in the post: R bloggers Bluesky bot using atrrr and uses tidyrss to read the RSS feeds.

You’ll need to install the packages and then adapt the script accordingly.

When you get to part 3 of the script, you need to run the function auth() from atrrr with the Bluesky name of the account you set-up in Part 1 and follow the prompts to get the app password that you will save for Bluesky authentication.

So for example I ran:

auth("protpapers.bsky.social")

This opens Bluesky and in settings I went to App Password, clicked Add App Password which created a random name for the password (e.g. PrussianBlue) and then I clicked to Create App Password to actually get the password associated with the name.

Paste the password back into the prompt in R and also save it in your .REnviron like so

Open .Renviron with the usethis package:

usethis::edit_r_environ()

Add the password like so:

ATR_PW="your-password-string"

The BSKY_TOKEN below was created when we used auth() and setting it here saves it to the papers_token.rds file. But you can call this whatever you want.

ATR_PW is then called into the environment like so:

Sys.setenv(BSKY_TOKEN = "papers_token.rds")
pw <- Sys.getenv("ATR_PW")

auth(user = "protpapers.bsky.social",
     password = pw,
     overwrite = TRUE)

You can always run auth() again and created a new password and token in the future if you are have problems.

Make sure your .gitignore is set not to upload files with passwords or other sensitive information e.g. your .Renivron file.

You can test the bot.R script first without deploying bot.yml on Github.

Adapt the YML for Github Workflow

To automate the bot you need to set-up a Github Action using the bot.yml.

The bot.yml script should work as-is as long as it’s in the .github/workflows/ folder.

I’ve set it run twice a day at 0525 and 1725 UTC (Github runs on UTC), but if you want to change the cron job, https://crontab.guru will help you with cron syntax.

With help from Claude, I’ve revised the .yml to speed it up using caching for system requirements and R packages. It should also catch errors and report them as log files too. Bot runtime is now about 1 min. 45 seconds.

It uses renv for caching and I created a GitHub Token for any parts that require authentication including R package installation. This is enabled by adding write permissions in the Github Action.

We need to add the password as a secret called ATR_PW in the bot repository to Github as repository secret in Settings > Security > Secrets and variables > Actions like so:

Everytime you push an update to Github, this will also trigger the workflow, otherwise it will run according to the cron times. Click on Actions at the top of your repo to see pending workflows.

@jbgruber.bsky.social added a Keep Alive job by https://github.com/gautamkrishnar, to the YML to prevent Github disabling the botif no commits are made to the repo after 60 days: https://github.com/marketplace/actions/keepalive-workflow.