Discover and read the best of Twitter Threads about #alphafold

Most recents (24)

We made #AlphaFold dream of new protein assemblies, used #ProteinMPNN to bring it back to reality, and the results were stunning! (1/5)…
Using #AlphaFold hallucination, we generated protein assemblies with up to C42 symmetry and validated 10 designs (7xtal/3cryoEM) – these structure differ significantly from anything the network has seen during training! (2/5)
What does this image and these proteins have in common? Activation maximization initially lead to adversarial sequences (insoluble), but #ProteinMPNN gave these structures a new life in the wetlab! (image: (3/5)
Read 5 tweets
1/ You’ve sequenced your samples and identified variants. Great! 🙌

Now, here's how you can use @ensembl to find out the effect of your variants on #protein structures and interactions. A thread...🧵

#genomics #bioinformatics #tweetorial #Ensembltraining 🧬 A screenshot from Ensembl's Alphafold protein viewer with va
2/ The Variant Effect Predictor (VEP) is what you’re going to need for this task. You can find it in the blue header from the @ensembl homepage:…

#VEP Screenshot of Ensembl's homepage with annotated arrow pointi
3/ Click the ‘Launch VEP’ button to open the VEP web tool and enter your input data using instructions in the documentation:
👉… Screenshot of Ensembl's VEP web interface documentation and
Read 10 tweets
More than 100 studies citing @Google 's #AlphaFold are being released each month according to the New Scientist. The protein structures and open-source tools previously released by @DeepMind have already accelerated many areas of research.
@Google @DeepMind Interestingly, the way AlphaFold models proteins resemble the way NLP algorithms make sense of sentence structures, as you can see from the chart in the head tweet above.
@Google @DeepMind I invested in a few AI companies over the last couple of years and continue to think that it's key to continue backing infrastructure players like DeepMind was in the early 2010s.
Read 5 tweets
Some perspective on #AlphaFold with the rather eye-popping announcement from @DeepMind and @emblebi (which I am Director of) of all known proteins (~200 million) having an AlphaFold prediction run.
First off at some conceptual level this is "just" about the scaleability of computational methods - if you can do 1 in a computer, you can do 200 million. But still, instantiating this resource in a systematic manner makes it real. But this instantiation is not trivial.
There are three types of engineering needed here; one is knowing, tracking and organising all known proteins - this is mission of @uniprot, a joint project of @emblebi, @sib and PIR of @Georgetown. Conceptually looks pretty simple but ... bamboozling detail, reconcilation, etc
Read 33 tweets
1/6 Can #AlphaFold learn from + & - labeled data to predict peptide binding better? w/ Phil Bradley @JustasDauparas @minkbaek @Mohamad_Abedi David Baker we present joint structure prediction-classification fine-tuning of #AF2 to classify pMHC binding.@UWproteindesign @fredhutch
2/6 We developed a template-based method for accurately predicting pMHC structures & fed confidence scores to a simple classifier to fine-tune with structure+classification loss on a self-distillation set of + and - pMHC interactions & reach SOTA in Class I & II classification.
3/6 We show that our method is potentially orthogonal to NetMHCpan and can correctly label peptides where NetMHCpan fails.
Read 6 tweets
In the recent years @DeepMind has been working on Protein Folding and thanks to their deep learning architecture AlphaFold they already managed to predict millions of proteins, accelerating the making of new protein-based drugs for fighting diseases.  🧵

#DeepMind #AlphaFold #AI Image
When mutated, normal cells might become cancerous. AlphaFold may determine how a mutation can alter the structure, and hence the function, of the proteins encoded by these genes. Understanding how mutations influence their structure can be key to developing new medical treatments
AlphaFold, is able to predict the structure of the protein in which a mutation occurs, see where that mutation is, and how the mutated protein interacts with other molecules, and ultimately get an idea about the mechanism, or why that mutation causes a problem for that cell
Read 3 tweets
1/ My new home @future just turned 1 🎉! Some thoughts on why I’m so excited about Future, by way of sharing a few of my favorite articles we’ve published so far. 🧵
2/ My colleagues share my optimistic outlook but get that building a better future won’t be easy. Many of my favorites tap into a common theme: science is integral to creating a better world, only if we (scientists) embrace new technologies and explore radical ideas, e.g.
3/ Science is slow, but it doesn’t have to be. One of the things slowing science down is the archaic process of funding science. Begging for #grants takes huge chunks of researcher time, decisions take months, and decision-makers are often risk-averse.
Read 11 tweets
Today is a BFD triumph in life science—solving the 3D structure at near atomic level resolution, one of the world's hardest, giant jigsaw puzzles—the nuclear pore complex—the largest molecular machine in human cells, with structure-based AI prediction
Extraordinary work and perseverance by multiple groups…
This wouldn't have happened without #AI #AlphaFold
Read 4 tweets
A "negative" result, but phenomenal thought piece from @naturalantibody / @antibodymap. TLDR predicting antibody-antigen interactions is pretty darn hard (1/5) #antibodies #machinelearning #alphafold…
Predicting Ab-Ag interactions is a sub-problem of the protein-protein interaction problem. There are many facets to consider here, including but not limited to, identifying the correct antigen (let alone the correct epitope), the correct paratope, orientation, etc (2/5)
@antibodymap's team show first that true Ab-Ag pairs (i.e. those where we know the Ab binds antigen) and false Ab-Ag pairs (i.e. Ag was randomly given to an Ab), the pIDDT scores are incomparable, suggesting score-based discrimination is HARD. (3/5)
Read 6 tweets
2/ #AlphaFold can predict the folded protein shape, but not the complicated folding process itself. Often cells use protein complexes called chaperonins to help other proteins to fold. Here we purified a chaperonin called TRiC from cultured human cells.
3/ Purification was possible only by using #CRISPR genetic scissors. We introduced a purification handle in one of the 8 TRiC subunits. This allowed purification of the endogenous chaperonin as opposed to an engineered version of it. Quite a TRiCK! #samplerevolution
Read 8 tweets
Congratulations to @demishassabis, John Jumper, and David Baker who will be awarded the Wiley Foundation 20th annual Wiley Prize in Biomedical Sciences on April 1: 🧵 1/7
Demis and John accept the award on behalf of the @DeepMind team who worked on #AlphaFold, which was recognised as a solution to the “protein folding problem” at CASP14 in Nov 2020: 2/7
From the start, we committed to giving broad access to our work and, in July 2021, we published our methods in @Nature along with the open source code.

Open Source:… 3/7 Image
Read 7 tweets
Systematically overlaying likely pathogenic variants over @DeepMind #AlphaFold predicted structures provides a unique opportunity to dive into the mechanisms of disease.

Here, NOD2 inflammatory disease-causing variants on the experimentally unresolved structure. 1/5
Data-wise, the @OpenTargets Platform and Genetics Portal provide a comprehensive aggregation of disease-causing variants post-processed from other great resources like @GWASCatalog or @NCBI_Clinical ClinVar (via @evarchive) among others…
Now we also make human #AlphaFold structures available as a result of the collaboration between @embl and @DeepMind. All thanks to the great protein annotation tools provided by @uniprot

See NOD2 platform profile page…
Read 5 tweets
Since the publication of #AlphaFold2 and #RoseTTAFold and now that the tools and models have been made accessible, there has been an avalanche of attempts to solve old crystal structures. This thread covers tips from the #PhaserTeam for doing #MR with these models. (1/...)
The first thing to be aware of is that the B-factor fields contain measures of confidence in the correctness of the prediction, not actual B-factors. This means: 1) we can use that confidence to trim the model and 2) we need to convert that to an appropriate B-factor. (2/...)
Phaser will take those B-factors and use them to weight the different parts of the model. This can improve your chances of success with the model. (3/...)
Read 11 tweets
Very excited to share that our work on Protein Nonrefoldability is out today in @J_A_C_S! #AlphaFold might be great at finding proteins' native structures... but turns out many proteins themselves are not! Short 🧵1/7…
We developed a mass spec approach to probe the refoldability of the proteome. First we unfold & refold E. coli extracts, then use a protease to interrogate the structures of 'refolded' proteins. The resulting peptide fragments are sequenced by LC-MS & compared to native. 2/7
We found that most simple 'model' proteins are well-behaved and can refold on their own. But lots of multi-subunit assemblies, multi-domain proteins, and certain fold-types cannot fully refold intrinsically. 3/7
Read 7 tweets
Now that the #alphafold hype has completely died down (ha!), I've written a new blog post on the AF2 method paper:…. This is a technical deep-dive into aspects of AF2 that I find most surprising/innovative and of relevance to broader biomolecular modeling.
My post is _not_ a high-level summary of how AF2 works. For that I suggest @c_outeiral's blog post….
Should say that we will have in a couple of weeks a formal review paper out that is a high-level overview of AF2 and its implications.
Read 3 tweets
Yesterday we announced early collaborations using the #AlphaFold Protein Structure Database, which offers the most complete and accurate picture of the human proteome to date. So how is AlphaFold helping these organisations with their work…? 1/ Image
The Drugs for Neglected Diseases initiative (@DNDi) has advanced their research into life-saving cures for diseases that disproportionately affect the poorer parts of the world. 2/ Image
The @CEI_UoP is using #AlphaFold's predictions to help engineer faster enzymes for recycling some of our most polluting single-use plastics. 3/ Image
Read 4 tweets
A personal view point on the #AlphaFold announcement today from the @DeepMind and @emblebi team, part of @embl. TL;DR - I am *still* pinching myself about this.
When @demishassabis and the AlphaFold team first presented the results from CASP to me last November I genuinely almost fell off my chair. I think I swore quite a bit (in a British way) in amazement.
One of the reasons was I knew how rigorous CASP was - 20 years ago people published all sorts of "solving the folding problem" which then... didn't work beyond the training set. CASP cleverly used the fact that there are genuinely unknown structures each year solved by experiment
Read 15 tweets
Thoughts on the #AlphaFold #Deepmind 'we done the proteome' news:
(1) This is great.
(2) This was always going to happen. I'm surprised they did it this fast, but cool.
(3) The dataset will be invaluable for hypothesis generation.
(4) Hypotheses will still need to be proven at the lab bench.
(5) This will advance structural biology, allowing phasing of MX datasets and tracing of domains/proteins into EM maps.
(6) I'm interested to see how their models perform with drug & biologic design.
(7) I'm interested to see how their model aid in construct design. Knowing disordered/unstructured regions is super important for designing well behaved, soluble recombinant proteins for structural study. #Alphafold
Read 18 tweets
Today with @emblebi, we're launching the #AlphaFold Protein Structure Database, which offers the most complete and accurate picture of the human proteome, doubling humanity’s accumulated knowledge of high-accuracy human protein structures - for free: 1/
We’re also sharing the proteomes of 20 other biologically-significant organisms, totalling over 350k structures. Soon we plan to expand to over 100 million, covering almost every sequenced protein known to science & the @uniprot reference database. 2/
We’re excited to see how this will enable and accelerate research for scientists around the world. We've already seen promising signals from early collaborators using #AlphaFold in their own work, including @DNDi, @CEI_UoP, @UCSF & @CUBoulder:… 3/
Read 5 tweets
Daily Bookmarks to GAVNet 07/16/2021…
How Many Numbers Exist? Infinity Proof Moves Math Closer to an Answer.…

#numbers #mathematics #InfinityProof
Read 8 tweets
Yesterday we shared the news that #AlphaFold has been recognised as a solution to the ‘protein folding problem’ by #CASP14, the biennial Critical Assessment of Protein Structure Prediction. But what exactly is protein folding, and why is it important? A thread… (1/6)
Proteins are the building blocks of life - they underpin the biological processes in every living thing. If you could unravel a protein you would see that it’s like a string of beads made of a sequence of different chemicals known as amino acids. (2/6)
Interactions between these amino acids make the protein fold, as it finds its shape out of almost limitless possibilities. For decades, scientists have been trying to find a method to reliably determine a protein’s structure just from its sequence of amino acids. (3/6)
Read 6 tweets
#Alphafold by #deepmind used solid interdisciplinary intuitions for algorithm/model design. It wasn't just a rinse-and-repeat machine learning exercise. Details on methods are limited, but here's my best interpretation (+some predictions) so far: [1/n]
Protein sequence databases provide us samples that have defacto passed the fitness test of evolution and are information-rich. "Genetics search" is a retrieval step to find nearest-neighbors as defined by sequence alignment. Why do we need nearest-neighbors (NNs), you ask?
There's a neat principle/intuition called coevolution that can help explain. The mutational variance observed can give clues to protein structure and function. Read more here:
Read 13 tweets
Thrilled to announce our first major breakthrough in applying AI to a grand challenge in science. #AlphaFold has been validated as a solution to the ‘protein folding problem’ & we hope it will have a big impact on disease understanding and drug discovery:
The ultimate vision behind @DeepMind has always been to build AI and then use it to help further our knowledge about the world around us by accelerating the pace of scientific discovery. For us #AlphaFold represents an exciting first proof point of that thesis.
Congratulations to the whole #AlphaFold team! And thanks to John Moult and the fantastic CASP organisers and community for championing this critical problem and creating such an amazing benchmark - it really is the gold standard for scientific assessment. #CASP14
Read 3 tweets
In a major scientific breakthrough, the latest version of #AlphaFold has been recognised as a solution to one of biology's grand challenges - the “protein folding problem”. It was validated today at #CASP14, the biennial Critical Assessment of protein Structure Prediction (1/3)
CASP is both the gold standard for assessing predictive techniques and a unique global community built on shared endeavour. Accuracy is measured on a range of 0-100 “GDT”. #AlphaFold has a median score of 92.4 GDT across all targets - its average error about the width of an atom.
We’re excited about the potential impact #AlphaFold may have on the future of biological research and scientific discovery. Thank you to the CASP organisers & the whole community - we look forward to the many years of hard work and discovery ahead:
Read 3 tweets

Related hashtags

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!