Simply Biotech’s Post

Gene expression is the process through which information inscribed into a gene is converted into a function. ML genetic and healthcare data specialist Dr Sara Moein relates how ML accurately predicts target genes from landmark ones and states the implications of this in medicine. More at https://lnkd.in/g72_h2e7. #ML #healthcare #data #gene #medicine #RNA

AAV has limited gene capacity. But this isn't the 90s anymore. Here, the authors cover 5 techniques to deliver large transgenes with AAV.   1. Minigenes and Functional Variants In this case, smaller is just as good. - Shortened genes that still do the same thing - Bioinformatics makes sure protein is still functional - Successful Examples: Mini-otoferlins and mini-PCDH15 show therapeutic potential   2. Expression Cassette Optimization Those non-transgene sections must earn their keep! - Promoter design can now be tailored for specific cell or tissues - ITRs can be modified to enhance transduction efficiency and expression. - Regulatory Elements can be custom designed to each specific transgene, boosting expression   3. Exon Skipping Cut out everything that's non-essential! - Used by cells to skip over problematic exons, leading to a truncated (but functional) protein - Can be used on larger transgenes to bring down their size to AAV's capacity - This method has been used to make CRISPR small enough to fit inside AAV   4. Multiple AAV Vectors for Large Genes One AAV isn't enough? Use 2 or even 3! Here's how you make it work: - Trans-splicing: Connects split genes for full-length protein. - Overlapping Sequences: Facilitates homologous recombination. - Inteins: Protein elements that enable post-translational splicing   5. Circular Permutation for Minigenes This is the (relatively) new kid on the block. - Used to test alternative combinations of the same protein - Can screen out non-essential components while maintaining functionality   Expect these techniques to play an out-sized role in AAV gene therapy for years to come.   Kudos to the authors, this is a must-read overview! Link for the open-access article is here (PDF wouldn't upload for some reason) https://lnkd.in/gJnBRyRk Any thoughts on this? Drop them in the comments.   For more papers related to viral vector process development, check out #pdjournalclub   #genetherapy #biotechnology #innovation #science

New #AI method, #LINGER, cracks the code on gene regulation! 💡 LINGER, developed by Clemson University scientists, infers gene regulatory networks from single-cell multiomics data by integrating atlas-scale bulk data and prior knowledge. It outperforms existing methods and enables interpreting disease variants from expression data alone. Quick Read: https://lnkd.in/gYViayW2 #bioinformatics #genomics #machinelearning #generegulation #bigdata #sciencenews #biotechnology

New #AI method, #LINGER, cracks the code on gene regulation! 💡 LINGER, developed by Clemson University scientists, infers gene regulatory networks from single-cell multiomics data by integrating atlas-scale bulk data and prior knowledge. It outperforms existing methods and enables interpreting disease variants from expression data alone. Quick Read: https://lnkd.in/gYViayW2 #bioinformatics #genomics #machinelearning #generegulation #bigdata #sciencenews

AAV has limited gene capacity. But this isn't the 90s anymore. Here, the authors cover 5 techniques to deliver large transgenes with AAV. 1. Minigenes and Functional Variants In this case, smaller is just as good. - Shortened genes that still do the same thing - Bioinformatics makes sure protein is still functional - Successful Examples: Mini-otoferlins and mini-PCDH15 show therapeutic potential. 2. Expression Cassette Optimization Those non-transgene sections must earn their keep! - Promoter design can now be tailored for specific cell or tissues - ITRs can be modified to enhance transduction efficiency and expression. - Regulatory Elements can be custom designed to each specific transgene, boosting expression 3. Exon Skipping Cut out everything that's non-essential! - Used by cells to skip over problematic exons, leading to a truncated (but functional) protein - Can be used on larger transgenes to bring down their size to AAV's capacity - This method has been used to make CRISPR small enough to fit inside AAV 4. Multiple AAV Vectors for Large Genes One AAV isn't enough? Use 2 or even 3! Here's how you make it work: - Trans-splicing: Connects split genes for full-length protein. - Overlapping Sequences: Facilitates homologous recombination. - Inteins: Protein elements that enable post-translational splicing 5. Circular Permutation for Minigenes This is the (relatively) new kid on the block. - Used to test alternative combinations of the same protein - Can screen out non-essential components while maintaining functionality Expect these techniques to play an out-sized role in AAV gene therapy for years to come. Kudos to the authors, this is a must-read overview! For more papers related to viral vector process development, check out #pdjournalclub #genetherapy #biotechnology #innovation #science

Recent study in Nature Biotechnology presents LINGER, a machine learning method that significantly improves the inference of gene regulatory networks by integrating single-cell multiome data with large external data. Conventional methods for inferring gene regulatory networks (GRNs) have predominantly used gene expression data or bulk data with lower resolution. Q. Yuan & Z. Duren in their research paper present LINGER (Lifelong Neural Network for Gene Regulation), a novel machine learning method that significantly enhances the accuracy of GRNs and thus the understanding of how genes are regulated in our cells. LINGER uniquely integrates single-cell gene expression and chromatin accessibility data with external atlas-scale bulk data. It also incorporates prior knowledge of transcription factor motifs to improve learning accuracy and depth. The study reports a 4 to 7-fold increase in inference accuracy compared to existing methods. LINGER not only maps a more detailed regulatory landscape, but also provides a new lens through which we can interpret disease-associated variants and genes. This is crucial to improve our understanding of genetic contributions to disease and to tailor medical treatments to individual genetic profiles. Read the full research article by Q. Yuan & Z. Duren, published in Nature Biotechnology on 12 April 2024, here: https://lnkd.in/eiytSjrv #Biotechnology #GeneExpression #GeneRegulation #GeneRegulatoryNetwork #GRN #LINGER #MachineLearning #ML #Multiomics #NatureBiotechnology #Omics #Research #SingleCell #WeeklyPublication

The 3 biggest mistakes we keep making in clinical operations: As we enter a new era of modern science with striking technological advancements from gene therapy to artificial intelligence (#AI), the industry is still running in place – making the same mistakes in clinical research, including: Mistake #1 – failure to recruit a representative patient population Mistake #2 – delayed response to operational problems Mistake #3 – reliance on the same, biased data sources (leading to various negative outcomes) https://lnkd.in/e6XT4Mji

IS THIS SUCCESS❓ This recent gene editing study has made the global news, and many friends sent it my way because it relates to Vicky’s condition (but a different gene). The scientific progress is absolutely incredible. The teams at Mass Eye and Ear, Oregon and Philadelphia did an amazing job, many of whom I personally know, and I admire. I’m happy for the patients who took part in this trial and saw improvements in their vision. When started, it was the very first gene editing done directly on the human body —ophthalmology leading the way again! But here’s the thing: this is only partly a success story. The study completed almost two years ago and just got published. What’s often missed in the headlines is that the work is currently paused. Yes, some people saw big improvements, and CRISPR is promising. But this wasn't enough to demonstrate a commercial return, leading to a lack of investment in further clinical development to ultimately make it available to patients. There remains no treatment for conditions like this (nor for Vicky's one), and this work is simply not progressing (yet). I absolutely get the P&L challenges and investor expectations. But this is another example of how traditional commercial models are lagging behind scientific advancements. We often shy away from small patient populations and rare diseases, failing to see that addressing these can pave the way for a (near) future of personalized medicine. We need to rethink how we bring these scientific advancements to patients. We need, among others: 🏥Better policies that encourage investments in rare disease and not hamper them 🚩Better regulatory frameworks to address the challenges of small patient population trials 👫Better partnership between industry and patients 🧫More efficient manufacturing to reduce the cost What else? Personalized medicine is here. We just need to find a way to access it… Link to the study https://lnkd.in/eQMqStTk #MyOwnViews #PersonalizedMedicine

Hi, my LinkedIn connections, I'd like to share with you an article from Nature that I read this week. 📌 My three key takeaways - Innovative Modeling of Genomic Contexts The genomic language model (gLM) leverages millions of metagenomic scaffolds to learn the functional and regulatory relationships between genes. By training on diverse genomic sequences, gLM effectively captures the complex dynamics of protein co-regulation and function, providing insights that are biologically significant and extend beyond traditional gene sequence analysis. - Contextualized Protein Embeddings and Function Prediction The gLM employs contextualized protein embeddings that encapsulate not only the protein sequence but also the surrounding genomic context. This method allows for a more nuanced understanding of protein functions and their interactions within the genomic framework. The model's ability to predict gene function and co-regulation based on these embeddings shows promising applications in computational biology and genomics. - Potential for Broad Applications The paper underscores the potential of gLM to be adapted for various applications, such as predicting enzyme function, discovering new biomolecular interactions, and enhancing the understanding of genomic structures. The model's flexibility and effectiveness suggest it could significantly impact how genomic data is interpreted and used in research, offering new avenues for discoveries in gene therapy, drug development, and more

AI-designed gene editing tools successfully modify human DNA https://buff.ly/3JzfsPZ #geneediting #therapeutics #biotechnology #biotech