We are excited to share our work assessing the impact of 1000+ industrial and agricultural pollutants on 22 gut microbiome species 🧪🧫🧬. A thread 🧵 Anna Lindell StephanKamrad Kiran Raosaheb Patil MRC Toxicology Unit doi.org/10.1101/2024.0… (1/9)

Don't forget to apply for one of our LMS Mathematics in Africa Grants, where a maximum of £2,000 partial funding is available for any maths activities (eg. events, conferences, research) taking place in Africa. Next deadline: 15 September. ➡️ lms.ac.uk/grants/mathema…

"why doesn't mankind build wonders anymore?" bro we built a cathedral to ignite the flame that burns at the heart of distant stars you're complaining that we spend too much money on it

CSV-Filter: a deep learning-based comprehensive structural variant filtering method for both short and long reads academic.oup.com/bioinformatics…

LevSeq: Rapid Generation of Sequence-Function Data for Directed Evolution and Machine Learning biorxiv.org/content/10.110…

μFormer: Accelerating protein engineering with fitness landscape modeling and reinforcement learning biorxiv.org/content/10.110…

Is progress in biology too slow? -- I tried to answer this question and ended up explaining why biotechnologists, economists and members of the general public might have different answers to this question writingruxandrabio.com/p/is-progress-…

Stardust. (Created with Midjourney v6.1)

Crystal Skies & HALIENE / “Stardust” youtu.be/wt20g8E42QI?si…

DeepKINET: a deep generative model for estimating single-cell RNA splicing and degradation rates genomebiology.biomedcentral.com/articles/10.11…

Oh, yeah, looks like I'm a professor now

Can we all have a moment of sympathy for DNA – it has had a rough week 🧬 First, proteins not predicted by genetic code 😮 biorxiv.org/content/10.110… Then, inheritance without genes through amyloid protein 🤯 nature.com/articles/s4155… Biology is amazing!

Analysis of biobanks is challenged by missing values of phenotypes. Our new paper on using ML-based synthetic surrogate analysis (SynSurr) to improve power for GWAS of partially missing phenotypes in population biobanks. nature.com/articles/s4158… (1/)

Executor. (Created with Midjourney v6.1)

NY instagram.com/reel/C_irGQ7NZ…

GenoM7GNet, an efficient deep learning prediction model utilizing a nucleotide language model. GenoM7GNet primarily comprises two parts: a pre-trained Bidirectional Encoder Representation from Transform (BERT) model and a CNN model.

GenoM7GNet utilizes DNABERT model on human genomic data as an embedding layer to embed tokens into real-valued vectors. GenoM7GNet employs a one-dimensional CNN to learn the vectors outputted from the BERT embedding layer, thereby achieving the identification of m7G sites.