AMPSphere published in
@CellCellPress
. A thread on the main results
We obtain nearly 1 million novel antimicrobial peptides from the global microbiome!
In vitro testing showed that at least 79/100 are true positives
So far, when it comes to LK-99, everyone is behaving like the cliché:
China: a bunch of university labs worked to replicate as fast as possible
US: a private company sponsored a replication as a marketing gimmick
Russia: cat girl, but whole thing is unclear
Europe: on holiday
Imagine unscheduled phone calls didn't exist and someone came up with the idea:
"it's an alarm to schedule a meeting RIGHT NOW."
"And who can ring the alarm?"
"That's the beautiful part: anyone in the world can ring a loud alarm in your house or office!"
"I'll pass, thanks."
"Towards the biogeography of prokaryotic genes" is finally out at
@Nature
We wanted to know what genes are out there and how they are distributed around the world
A 🧵 on the main findings
1/18
It's Nobel week, so—after a break last year when this account became a Karikó stan account—I am reposting my annual plea to give Lipman and Myers a Nobel Prize for creating BLAST
@btharris93
If anything, there is a stronger signaling/deterrence argument for throwing her in jail
Sending a message that "spying for a foreign power is not okay" is still relevant in a way that "leading an SS Einsatzgruppe is not okay" is not
@timminglab
Emeritus Professor: I was supposed to start this talk 7 minutes ago, but I forgot about it until someone came to get me in my office. Let's do this!
Remember when people would pretend that monkeypox spread had nothing to do with sexual contact?
That we were just on the verge of seeing the disease spread outside MSM networks?
Why are bioinformatics results so full of false positives?
I've been thinking about this for a few years: the incentives in the field are to produce false positives
(link in next tweet)
@apenwarr
It's good UX to check if a file exists as soon as you are aware that it will be needed
A pipeline that runs for 2 hours before erroring out because the user mistyped an input file that is only needed at step 5 is bad
(Agree on check-before-opening being wrong, though)
@Hieraaetus
European states will send someone with a gun to arrest you if you shoot a wolf. If you resist arrest, they will send more people with more guns
They have a stronger hold on the monopoly of violence so they often don't need to use as much violence, but the threat is there
@Finnothyjest
Europeans do this all the time: get very upset that conservatives in the US are doing something to make the US more like Europe
Chevron is a bit too niche, but certainly Dobbs and immigration fall into that category
@JosephPolitano
This is, at minimum, misleading
Direct election of Head of State is common, but direct election of head of the executive (most often, prime minister) is not
Both "we don't want to pay full tuition for Zoom classes" and "we are working extra to provide this service" are 100% correct and not contradictory
The simple fact is that covid has made the world poorer
This is what it means to be poorer: we work harder and get less
Perception: Using an institutional email in a paper affiliation: Professional
Using a gmail/personal email: Unprofessional
Fact: My personal email has been the same for 15 years, during which time I had 7 institutional emails, so all my old papers have unusable addresses
As a reviewer, I think that bioinformatics/compbio papers can make three different types of contribution:
1. A new method/algorithm/insight (bioinformatics advance)
2. Novel biological insight (computational biology)
3. A tool for others to use (public goods contribution)
@AlexGodofsky
In Europe, micro filtration is allowed as an alternative to pasteurization as it achieves similar levels of bacterial load reduction
If this is what they are selling, it's a good product and what I used to consume in Luxembourg
Woohooo!! My
@arc_gov_au
Future Fellowship application got funded!
We'll be working on the small proteins of the global microbiome!
I'm very excited about this science and I'm very happy that
@arc_gov_au
agreed with me on the promise of small proteins
@GarrettPetersen
@Aella_Girl
Bacteria already evolved to digest plastic on their own, but it is not a great energy source, so there is actually little risk that they will start digesting the water bottle on your cupboard shelf
@Noahpinion
It's a general cliché that it's the diaspora that takes irredentist positions, while the locals want a peace treaty so they can get on with their lives
The Palestinians have an extended diaspora, fighting for their cause, but not for them
Internet punditry, circa 2011: you try to blog your way into a journalism job that actually pays the bills
Internet punditry, circa 2021: you try to journalism your way into a blog that actually pays the bills (we call them newsletters now)
- Tents in church, mall, school parking lots 7 days/week
- Staff to the gills, workers get hazard pay
- Allot a certain number of doses daily
- 6 am - 6 pm: HCWs and 60+ only
- 6-10 pm: remaining doses open to all
Seems doable with 1920s-level US logistical competence.
I used to come home from conferences with a long list of research ideas. Both completely new projects as well as minor tweaks to ongoing projects
Many of these were bad, mind you, but the interactions were very stimulating intellectually. Zoom conferences don't do this
New experiment: I will work on a project (starting question: what antibiotic resistance genes co-occur in genomes/metagenomes and does it matter?) completely in the open
Dear submitter of papers in biology with neural network results, please include a baseline using logistic regression to show some value to all that computing
Signed, reviewer
#2
The theory that that, if China hadn't delayed in admitting how bad things were, we would have prepared makes perfect sense, except for the existence of February
In February, it was clear it was bad and little was done — it even lasted 29 days this year
Good news : Germany is lowering its CO2 emissions in 2018 after years of stagnation or increasing emissions
Bad news: mostly because global warming makes people use less heating
Recently, I read "Lying for Money", which introduced me to the concept of "criminogenic environment"
I think many PIs create "fraudogenic environments" in their labs while keeping plausible deniability
A culture that "reward[s] 'winners' (postdocs who generate favorable results) and marginalizes 'losers' (postdocs who struggle to generate such data)” leads to 20 years of "fudging of results [in] labs at three separate institutions".
Flight attendant: Is there a doctor on board?
Dad: That could have been you... Go see if they need a bioinformatician!
FA: Bioinformatician? Perfect, the passenger in 12B needs help with bioconda and incompatibility errors!
Me, sighing: I cannot help him... Nobody can
Latest Stanford study shows again that serology surveys are NOT useful when the prevalence is known to be lower than the test's false positive rate
Here, the authors tried to argue that their test was specific enough, but they need more evidence for that
@Jess_Osterhout
Selling and networking are a part of the job, as much as anything else. You can and should get better at it
Written rules are often guidelines
Distinguish tasks that need your best (typically, your science) and those that don't (typically, internal tasks)
You know that reviewer that writes "this is a good paper, but belongs in a lower tier journal"?
I'm writing a review and am tempted to write "this is so good, it belongs in a higher tier journal."
How many times do I need to explain that I didn't sell my soul to the devil, I sold a token which is associated with the hash of my soul on the blockchain to the devil?
It's that time of year, when I repost my plea that BLAST get Its Nobel Prize. Not that I have any pull with the Academy, but I still think it should get one
Building MAGs (metagenome-assembled genomes) from multiple metagenomes using SemiBin
Using multiple samples improves results at the cost of more computation
@CutTheKnotMath
Here's the intuition: once Alice gets a first heads, she will keep tossing only while she gets only heads: the first tails and she is done. While Bob, if he sees a heads followed by a tails, he's back at the beginning.
What the rest of Europe should learn from Italy is not "we're all going to die". We're not
The lesson is "once you get past 100 new cases per day, shutdowns are inevitable. Waiting only means that they will last longer and more people will be dead at the end of it"
I was reviewing a paper today and explicitly tried to summon my inner
@MicrobiomDigest
to spot any duplications
I didn't see any, but before following her on Twitter, it would not even have crossed my mind to actively search for duplication
@StefanFSchubert
This plus the stagnation of Europe and Brexit means that, contrary to the narrative of the early 2000s, the US is becoming even more dominant
China is such a "high trust society"
In the last few weeks, I have sold off a few things by (1) posting a few pictures online, (2) having complete strangers text me money and their address, and (3) shipping them their things
Not even a reputation mechanism
The three levels of research code:
1) one-off script
2) generates plot/table for paper
3) library for reuse by others
There is a 5-10x effort difference between each level, so library code should be 25-100x more effort than a one-off
@AdamSinger
There is a collective action element here
If a parent helps a child, that helps the child. But if all parents helped their children, the effect would be different
We fundamentally need more housing
They: "When the facts change, I change my mind. What do you do, Sir?"
Me: "I question the processes that led me to be wrong in the first place. To simply change one's mind about the object level is not enough."
In August 2023, I will join
@CMR_QUT
in Brisbane (Australia) and I'm very excited to join the great team they have there!
My scientific interests remain the global microbiome (including small proteins), with a mix of method development and applications projects