@bensb@genomic.social
@bensb@genomic.social avatar

bensb

@bensb@genomic.social

Group leader at the Ludwig Institute for #Cancer Research in Oxford. We do computational #genomics and #bioinformatics with a particular focus on #epigenetics and #mutation accumulation in the lead up to cancer. Also working on #cfDNA and early cancer detection in general. Views my own.

#fedi22 <- for appearing in searches on fediverse.info

This profile is from a federated server and may be incomplete. Browse more on the original instance.

aram, to ai
@aram@aoir.social avatar

What fresh hell

@academicchatter

bensb,
@bensb@genomic.social avatar

@aram @academicchatter scite.ai is actually not a bad product, this email is really poorly written though. That's what happens when your sales team is just out for a quick bonus πŸ˜…

Basically, scite has a large corpus of citation statements in the context of the paper from which they came. They have a chat bot which will search their data and then use chatGPT to summarise the information extracted from their corpus. It often works quite well as a starting point for lit review etc

furqanshah, to academia

What do we want? Open science + transparent peer review!

When do we want it? Now!

Yet, reviewers will hide behind the cloak of anonymity. As an editor, there is little to be done about such behaviour. πŸ˜”

πŸ§ͺ

@academicchatter @academicsunite @ScienceCommunicator @openscience

bensb,
@bensb@genomic.social avatar

@furqanshah @academicchatter @academicsunite @ScienceCommunicator @openscience

Non-anonymous peer review creates a different, and more common, type of conflict of interest!

As a young PI without tenure, I'm reviewing papers by senior people in the field who wield power over my career, because they sit on all the grant and hiring committees, and will review my papers and proposals. If I cause their paper to get rejected from Nature, I do not want my name put next to the review, sorry.

cdamian, to techno
@cdamian@rls.social avatar

Aciiiiiiiiiiiiiid!

KI/KI pres. 5HRS OF ACID at Warehouse Elementenstraat

https://youtube.com/watch?v=VfcEAfmNVGM&si=ydgTi_vn3DTwiskA

> Relive two hours of the surreal acid set KI/KI played back in March/April 2023 at the notorious Warehouse Elementenstraat in Amsterdam.

bensb,
@bensb@genomic.social avatar

@cdamian always been more of a minimal person, but that set is wild

rwg, to academicchatter
@rwg@aoir.social avatar

Hey, @academicchatter, many folks are trying to figure out an alternative to Google Scholar.

Does anyone regularly use worldcat.org for that purpose? I'm trying it a bit and... it's working

bensb,
@bensb@genomic.social avatar

@rwg @academicchatter Another vote for semanticscholar, especially for alerts. They work brilliantly: I have a few folders with related papers, and every day I get an email with highly relevant literature for each of the topics.

kjschmid, to random

I just deleted my ResearchGate account, because πŸ‘‡

(Never used ResearchGate anyway, so no loss for me)

bensb,
@bensb@genomic.social avatar

@gpollara @kjschmid @academicchatter yeah, similar here. I think the business case for an "academic LinkedIn" was never very strong, so they are increasingly pivoting towards publication, trying to position themselves somewhere between scihub, biorxiv and the journals. I'm not sure they've really found the sweet spot yet, and this collaboration with MDPI doesn't bode well...

radlschorsch, to fediverse German

Wo bleiben die UniversitΓ€ten im Fediverse?

warum betreiben UniversitΓ€ten nicht schon lΓ€ngst eigene Mastodon-Instanzen?

Warum bekommen Studierende nicht mit der Inskription auch einen Mastodon-Handle?

Warum hosten UniversitΓ€ten Vorlesungsvideos nicht ΓΌber PeerTube im Fediverse?

Es gibt einiges, das fΓΌr ein stΓ€rkeres Engagement von UniversitΓ€ten im Fediverse spricht.

Ein Aufruf den man nur unterstΓΌtzen kann!

https://netzpolitik.org/2023/aufruf-hochschulen-aller-laender-ins-fediverse/

bensb,
@bensb@genomic.social avatar

@srfirehorseart @radlschorsch @academicchatter @edutooters I don't think this has to do with "government policy", but rather with a) inertia and b) avoiding the responsibility of content moderation. The latter, I think, is unfortunately a big one. Universities would feel exposed to legal and reputational risks for posts disseminated on "their" fedi instance.

fulelo, to Germany
@fulelo@journa.host avatar

- election: Toxic campaign heralds big vote for 's
https://www.bbc.co.uk/news/world-europe-67018486

bensb,
@bensb@genomic.social avatar

@fulelo yet more dreadful news on this wretched weekend πŸ˜“

ct_bergstrom, (edited ) to random

AITA, academic publishing edition.

Journal sends a review back to me because as reviewer I did not run the code and replicate the results.

My reply: "As an unpaid anonymous peer reviewer who handles probably 30 manuscripts a year, I am absolutely not in the position of running and evaluating code any more than I am in the position of running gels to evaluate lab results. If this is important to you, I suspect you can find a consultant who will do it at a reasonable rate."

Am I the asshole?

bensb,
@bensb@genomic.social avatar

@ct_bergstrom first off, I'm AMAZED that a journal asked you to do it. In my field (genomics) this never happened and from my experience the literature is riddled with missing data and erroneous code. I personally do try to run code and check raw data. This means I only review a handful of papers a year. So be it! It's not our job to deal with the excess of papers, our job is to keep the published record as "clean" as possible.

Private
bensb,
@bensb@genomic.social avatar

@clementaubert @ninokadic @academicchatter same. I'd leave it out unless there's something really impressive/related, like running a science podcast or something

Private
bensb,
@bensb@genomic.social avatar

@MartinVuilleme @academicchatter with genomic data, it is now 100% expected to upload all data to existing specialised repositories. They managed to upload some of the data, and claim in the papers that all data is available.The journal also requires that.

I don't suspect unreliable data, I just think they want to avoid competition and therefore withhold key data from potential competitors.

Asking to "collaborate" is not really an acceptable request if you've supposedly "published" the data.

bensb,
@bensb@genomic.social avatar

@Mehrad @academicchatter many years ago I went down that route. Took me almost a year to get them to finally tell me that to get the full data, we should have a "meeting to discuss collaborations" but that never materialised.

As a young academic at the time I shied away from openly confronting them, to prevent making powerful enemies πŸ€·β€β™‚οΈ so I just moved on. But now another young colleague told me that two more papers of them have the same issue, so I'm reconsidering moving to step two...

bensb,
@bensb@genomic.social avatar

@academicchatter I just noticed this on the Nature authors page. Basically, they now say they will chase authors down if they refuse to share. Might be worth a try...

bensb,
@bensb@genomic.social avatar

@sharris @academicchatter That's a good point! These guys are not in the UK, but very possible that their funding has similar requirements. Anyway, I think I'll go the way of contacting the journal(s) first, let's see where that leads...

ChristosArgyrop, to python

Until a truly performant (= fast, low memory footprint) two dimensional storage ("table") type (*) emerges, what are the options for managing big data in ?

  1. DBI into a performant DBMS (/ column store/ )
  2. shell over #R's data.table or 's / data.table packages, use files to get data in and some form of IPC to get data out
  3. , others ?
    (*) this a list of things one could encapsulate as objects for
    https://duckdb.org/2023/04/14/h2oai.html
    @Perl
bensb,
@bensb@genomic.social avatar

@ChristosArgyrop @Perl If you need this to work in Perl, then is an excellent solution. It is extremely fast and memory efficient (more so than data.table, since you have more explicit control of types). It can't do multi-type tables though, so in that case you need to write a lot more code to get what data.table gives you out of the box.

Edent, to random
@Edent@mastodon.social avatar

πŸ†• blog! β€œSometimes gzip beats Brotli”

Perhaps this was obvious to you, but it wasn't to me. So I'm sharing in the hope that you don't spend an evening trying to trick your webserver into doing something stupid. For years, HTTP content has been served with gzip compression (gz). It's basically the same sort of compression algorithm you get in a […]

πŸ‘€ Read more: https://shkspr.mobi/blog/2023/08/sometimes-gzip-beats-brotli/
βΈ»

bensb,
@bensb@genomic.social avatar

@Edent But how does the server know that gz will produce a smaller file, when - as I presume - the compression happens on the fly?! I've not run a static webserver in ages, so apologies for my ignorance...

Edent, to random
@Edent@mastodon.social avatar

All other things being equal, which of these coding conventions do you prefer?

property(object) or object.property() ?

For example, length($things) or $things.length()

What about if you are manipulating the object? E.g. upperCase($text) or $text.upperCase()

(Not doing a poll because I'd prefer a discussion.)

bensb,
@bensb@genomic.social avatar

@Edent IMO, an OO language should make it clear what constitutes a property of an object. As such, object.property is the right way to go.

BenjaminHCCarr, to Medicine
@BenjaminHCCarr@hachyderm.io avatar

Cancer cells-to-be accumulate a series of specific changes in a predictable and sequential way years before they are identifiable as pre-malignancies:
’s origin story features predictable plot line, researchers find
share
cells evolving in the laboratory undergo a series of predictable, sequential genetic changes that lead to pre-cancer. Blocking these changes may allow intervention before cancer occurs. https://med.stanford.edu/news/all-news/2023/07/pre-cancer-evolution.html

bensb,
@bensb@genomic.social avatar

@BenjaminHCCarr @fulelo Always great to see my line of research (in the widest sense) in the "news". However, I think the press release here goes a bit too far in suggesting that cancer generally develops in a sequential way. This experiment was highly artificial. Data from real patients' pre-cancerous lesions are very heterogeneous, suggesting there are many ways to form a tumour, not just one linear deterministic path!

villares, to random
@villares@ciberlandia.pt avatar

A great story via @Edent:

A customer complained that their bank was spelling their name incorrectly. The bank refused to solve the issue because it didn't have support for diacritical marks on US mainframe era software, so the customer raised a GDPR complaint under Article 16, sued and won!

"EBCDIC is incompatible with GDPR"
https://shkspr.mobi/blog/2021/10/ebcdic-is-incompatible-with-gdpr/

bensb,
@bensb@genomic.social avatar

@villares @Edent it's now 2023 and I can't use my correct name in most banking apps, nor with most airlines, nor at the UK government website etc πŸ€¦β€β™‚οΈ Many pages don't even accept the hyphen in my name! If I were a crypto billionaire with too much time and money I'd sue them all πŸ˜…

glynmoody, to random
@glynmoody@mastodon.social avatar

Twitter rival Mastodon rife with child-abuse material, study finds - https://www.washingtonpost.com/politics/2023/07/24/twitter-rival-mastodon-rife-with-child-abuse-material-study-finds/ the key point is: "A significant portion of the child abuse material researchers uncovered was from networks in Japan, where there are β€œsignificantly more lax laws"

bensb,
@bensb@genomic.social avatar

@glynmoody πŸ™ˆ "Telephone rival 'internet' still rife with child-abuse materia, study finds"

If someone doesn't understand the difference between a company and a protocol, then they seriously failed at being tech journalists

timkmak, (edited ) to random
@timkmak@journa.host avatar

Are you finding Mastodon to be less vibrant/active?

I’m worried that we are not getting as much engagement on our Ukraine war reporting as we used to.

We’re trying to decide which social media platforms to continue posting on, given how much effort it takes to replicate across all of them.

If you want to send a signal to us that we should keep posting here, will you sign up for our newsletter? http://Counteroffensive.substack.com

I’m only asking Mastodon users today so we can gauge impact.

bensb,
@bensb@genomic.social avatar

@timkmak Just discovered your account via this post. I won't subscribe to the newsletter, because my inbox is overflowing as is - but I really do read my mstdn timeline, and I am very interested in your content. Please do stick around here, if resources permit. I've been on Threads and find it appaling, so I'm hoping you'll continue to post to 🐘!

bensb, to random
@bensb@genomic.social avatar

guess the shroom πŸ™‚

image/jpeg

drmikepj, to random
@drmikepj@mastodon.social avatar

The longer I work in the British university sector, the more I'm in awe of anyone that achieves any kind of meaningful research output. It's almost as if we've created the perfect system for making people frustrated and burned out.

bensb,
@bensb@genomic.social avatar

@yaxu @drmikepj what's an FLF?

erictopol, to random
@erictopol@mstdn.social avatar

Not everyday you get a $282 million NIH award !

https://www.scripps.edu/news-and-events/press-room/2023/20230713-topol-all-of-us.html

bensb,
@bensb@genomic.social avatar

@erictopol congratulations!

EU_Commission, to random
@EU_Commission@social.network.europa.eu avatar

We reached a new milestone in ensuring the openness of digital markets.

Last week, 7 companies notified us they qualify as 'Gatekeepers' under the Digital Market Act.

We will check their submission before designating them. They will have 6 months to comply with the and, among others, be unable to:

  • lock in users in their ecosystem;
  • decide which apps you need pre-installed or which app store to use;
  • treat their own products and services more favourably.

More: https://europa.eu/!H4FQJM

bensb,
@bensb@genomic.social avatar

@EU_Commission @gruber so the law hasn't even taken effect yet - it'll be interesting to see what this leads to

gedeonm, to random
@gedeonm@mastodon.social avatar

I thought there were far more than a mere 1.2 million active Mastodon users.

bensb,
@bensb@genomic.social avatar

@gedeonm found it: https://mastodon.social/@Gargron/110675497622690306

  • All
  • Subscribed
  • Moderated
  • Favorites
  • β€’
  • JUstTest
  • ngwrru68w68
  • everett
  • InstantRegret
  • magazineikmin
  • thenastyranch
  • rosin
  • GTA5RPClips
  • Durango
  • Youngstown
  • slotface
  • khanakhh
  • kavyap
  • DreamBathrooms
  • provamag3
  • tacticalgear
  • osvaldo12
  • tester
  • cubers
  • cisconetworking
  • mdbf
  • ethstaker
  • modclub
  • Leos
  • anitta
  • normalnudes
  • megavids
  • lostlight
  • All magazines
    •