CindyWei

Bioinformatician, previous biologist, singer, theatre light designer, now into deeptech VC and precision medicine

[Mehran Karimzadeh @MKarimzade]

1/ Foundation models of the transcriptome, trained with generative pre-training or masked language modeling objectives, are confounded by gene co-expression networks. They shine at tasks like cell type and disease classification, but fail at downstream applications that

[Marija Kundakovic @KundakovicLab]

Extremely excited & proud to share our new preprint - we provide for the first time a single cell map of the mouse brain across sex, the estrous cycle, and peripartum! Data are 🔥 👇

"Single-cell map of the female brain across reproductive transitions"

https://t.co/zaqvE2CSW6

[Jonathan Gootenberg @jgooten]

Precise DNA integration is key for many therapies, but inserting large, scarless sequences remains challenging. What if we could reprogram nature's mobile elements for programmable genome writing?

Today we (+ @omarabudayyeh) debut Stitchr in @Nature

[Holger Heyn @hoheyn]

🚨 Inflammation Atlas – Version 2 is here! 🚨
MAJOR update of our INFLAMMATION ATLAS, including >6.5 million single cells from 1,047 patients across 19 diseases.

Inflammation touches nearly every aspect of human health—from infections to cancer to chronic immune-mediated conditions. Yet, a universal, holistic understanding of inflammation has remained out of reach. With this release, we move one step closer.

Using single-cell genomics and machine learning, we chart the full spectrum of immune cell activation in peripheral blood. The data, generated from a simple blood draw, opens the door for universal, non-invasive diagnostics that can identify and classify inflammatory states across diseases.

With this study we lay the groundwork for #PrecisionMedicine tools that can transform how we #diagnose and #treat inflammation-driven diseases. Let’s harness our immune system's signals to build a new era of #diagnostics.

🔬 Congratulations to all our partners @humancellatlas @cnag_eu @DoctisEU.

https://t.co/qrhCpVaZsf

[Thomas Norman @thenormanlab]

Can we recreate the transcriptional states seen in single-cell atlas projects? In our new preprint we try to do this using large-scale Perturb-seq experiments. Along the way we discuss off-target effects of CRISPR epigenetic editors and identify drivers of key fibroblast states.

[Peggy Pei-Chi Wei @PeiChiWei1]

Very important finding from the Glover lab on the role of PolQ in structure variation. They showed MiDaS is not responsible for SV under replication stress. Must read! https://t.co/to7QNWoc96

[Eric Topol @EricTopol]

Just out @Nature
The 1st whole-slide digital pathology #AI foundation model pre-trained on large-scale real-world data, from over 1.3 billion images, 30,000 patients
https://t.co/K3mKkHPdtu
@hoifungpoon @HanwenXu6 @Microsoft @UW @naotous @MSFTResearch

[Jian Ma @jmuiuc]

Excited to share scGHOST, now published @NatureMethods, graph-based #ML identifying #3Dgenome subcompartments in single cells. Kudos to @KyleXiongCMU & @RuochiZhang, who worked so closely on this. Exciting time for single-cell epigenomics & multiomics! https://t.co/R1VcdnCC3G

[David R. Liu @davidrliu]

Today in @Science_Magazine we report the laboratory evolution of compact degrons that enable targeted protein degradation triggered by an otherwise-inert small molecule, a multidisciplinary study that integrates organic chemistry, molecular glues, protein evolution, genome editing, and structural biology.
PDF: https://t.co/cCVLKL0w8Y
1/13

[Itai Yanai @ItaiYanai]

How did we lose our tail? A simple question.. but it wasn't really asked before! We discovered a plausible scenario for the genetic mechanism that led to tail loss. Amazing that such a big change may have been caused by such a small genetic event. https://t.co/0ZR8aH23PJ @BoXia7

[Niko McCarty. @NikoMcCarty]

Biology is (still) the world's most advanced technology.

Here are four interesting new papers, published yesterday.

***
Two new adenine base editors.
Papers: https://t.co/gOvcnQKPkN and
The first base editors were known as cytosine base editors because they could substitute a C --> T. Later, researchers made adenine base editors that can convert an A:T base pair, in a strand of DNA, into a G:C base pair.

Adenine base editors were originally made by fusing a wild-type TadA protein from E. coli, and an evolved version of TadA, to a Cas9 nickase. TadA is called an adenosine deaminase, because it deaminates an adenosine into an inosine molecule in tRNA molecules. A Cas9 nickase is just a Cas9 protein that 'nicks,' rather than cuts through, DNA.

Both of these new papers appeared in Nature Biotechnology. In the first, researchers evolved a new form of the TadA protein to make an adenine base editor that is able to cut more regions in the genome; in other words, they expanded the genomic "context" in which the protein works.

In one experiment, the adenine base editor was applied to human cells in culture that possessed a genetic mutation that causes Stargardt disease, which causes vision loss in children. One of the evolved adenine base editors, ABE8r, corrected the mutation in about 80% of the tested cells.

For the second study, researchers did away with TadA entirely! Instead, they made "deaminase-free base editors" by fusing a cytosine- or thymine-DNA glycosylase directly to Cas9 nickase. They then evolved these proteins, in the laboratory, to make them work better. The cytosine version of the base editor could convert C-to-A with an efficiency of about 60%.

A way to control protein expression levels in cells.
(Paper: https://t.co/W4MnYoyoGj)

In living cells, messenger RNA acts as a template that ribosomes read to build proteins.

This mRNA can get 'tagged' with various small chemicals, which make the mRNA more or less stable, or help coax ribosomes over to begin assembling a protein. The most common chemical modifier is called a, and it's important because it recruits proteins to mRNAs, and thousands of mRNA molecules are tagged with it.

Having too much or too few N6-methyladenosine tags on mRNA can be very bad. Dysregulation contributes to "cardiovascular disease, the response to viral infection and several cancers," according to this paper.

But until now, it was difficult to track which mRNAs are actually tagged with N6-methyladenosine in real-time. For this paper, researchers made a sensor "that provides a fluorescent readout when m6A is deposited on mRNA." This biosensor can also be adapted, the researchers showed, to actually throttle protein levels up-or-down.

In one example, a modified form of the biosensor was used to sense excess N6-methyladenosine in cancer cells, and then to dynamically slow down the growth of those cells.

An antibody for malaria.
(Paper: https://t.co/M3nK5SFFGu)

The first malaria vaccine recommended by the WHO for use in children, RTS,S/ASO1, has an efficacy in children of "45% after the first dose," but then "wanes to 36% over 4 years of follow-up," according to this new study.

An estimated 619,000 malaria-related deaths were reported in 2021. More antimalaria medicines are needed. (More reading: https://t.co/ZKoQuazDKR)

In this Nature Medicine paper, researchers sequenced the plasmablast repertoires (cells that make antibodies) from 45 people who had received the RTS,S/ASO1 vaccine in an effort to find antibodies that bind to malaria antigens. They generated more than 28,000 antibody sequences, tested 481 of them in vitro, and tested 125 of them in mouse models.

Two of the antibodies were selected for advancement, one was engineered so that it's easier to manufacture, and that engineered antibody was then advanced into clinical development. More hopeful news for malaria!

***
Thanks for reading. If you are interested in biotechnology, and would like to learn more, consider subscribing to Asimov Press. https://t.co/UGPAiIL7mw

[Hung Lo 羅鴻 @HungLo_tw]

Storytime! Please RT/share!
I am hyped to share that my PhD work is now preprinted (also my 1st-1st author paper). It was quite a journey!

https://t.co/dV6H4kIPPQ

It’s a story about binge eating and flavor, a short thread here: 1/n

[antisense. @razoralign]

The Interplay Between Sketching and Graph Generation Algorithms in Identifying Biologically Cohesive Cell-Populations in Single-Cell Data https://t.co/HRfSrToYh2

[Karsten Rippe @KarstenRippe]

Here is the free reprint link https://t.co/xZ3TX8y3wy of our @COSB_CRSB review on approaches for chromatin subcompartment analysis for the special issue "3D Genome Chromatin Organization and Regulation" https://t.co/XeVhV1wHUn

[antisense. @razoralign]

SCGP: Characterizing tissue structures from spatial omics with spatial cellular graph partition https://t.co/bHMGpqjuhE

[David van Dijk @david_van_dijk]

We are excited to present the Continuous Spatiotemporal Transformer (CST), a new transformer architecture that is designed for interpretable modeling of continuous dynamical systems such as PDEs and the brain🧠 https://t.co/JYpT5DJXax. Work with @aho_fonseca @E_Zappala @josueortc

[Jure Leskovec @jure]

📢 Exciting News! Our latest paper is now out in Nature Biotech 🌱🧬 We developed GEARS---an AI method to predict cellular responses to genetic perturbation. 🧪🔬

🔗 Link to the paper: https://t.co/pyhBTWrsGf

🧬 Unraveling genetic interactions in cancer, regenerative medicine, and more is a complex puzzle. GEARS harnesses the power of deep learning and a gene relationship knowledge graph to predict transcriptional responses to single and multigene perturbations using single-cell RNA-sequencing data. 🧠📊

🧪 What's unique about GEARS? It can predict outcomes for gene combinations that have never been experimentally perturbed before. 🚀

🔍 In a combinatorial perturbation screen, GEARS displayed a 40% higher accuracy compared to existing approaches. It identified distinct genetic interaction subtypes with remarkable accuracy and pinpointed the strongest interactions twice as effectively as previous approaches. 💥

📈 With GEARS, we're opening up new avenues for designing perturbational experiments and gaining deeper insights into the diverse effects of multigene perturbations. 🛠️

Join work with amazing @yusufroohani and @KexinHuang5.

[CompCancer @CompCancerRTG]

Do your PhD with us and apply now for fully-funded PhD positions in CompCancer! application deadline is 15th of August 2023, https://t.co/evvzPuqJwA

[Rafael Irizarry @rafalab]

Clustering algorithms report clusters even when none exist. In single-cell RNA-Seq pipelines, novel cell types are often identified by clustering algorithms. Expanding on Kimes et al.'s work, we introduce significance analysis for single-cell RNA-Seq data: https://t.co/ut1kPjKVCM

[Tirosh Lab @TiroshLab]

We’re excited to share our preprint on spatial organization in glioblastoma: https://t.co/OTaanCDDXG. We combine spatial transcriptomics (Visium) and proteomics (CODEX) to reveal a five-layer organization of cellular states and highlight a role of hypoxia as tissue organizer. 1/