Asimov Press的动态

Today at Asimov Press, we published a science fiction short story about the future of biology + computational models for drug discovery. Check it out at the link below!

1. Electrical synapses play a significant role in decision-making processes in animal brains, alongside chemical synapses. 2. Research conducted by Yale and the University of Connecticut, led by Dr. Daniel Colon-Ramos, highlights this role. 3. Animal brains filter vast amounts of sensory information to focus on critical details for survival. 4. Action selection refers to the brain's ability to prioritize context-relevant signals over irrelevant ones. 5. The study used the simple worm C. elegans to explore decision-making behaviors. 6. C. elegans exhibits advanced behaviors, such as navigating toward preferred temperature zones. 7. The worm employs gradient migration and isothermal tracking strategies to adjust behavior based on environmental context. 8. Electrical synapses, mediated by the protein INX-1, were identified as key in locomotion decision-making. 9. These synapses filter weak signals and prioritize significant sensory input. 10. Alterations in electrical synapses can drastically change an animal's behavioral choices. 11. Findings have broader implications, as electrical synapses are present in many species, including humans. 12. The study enhances understanding of sensory responses and decision-making, with potential applications in human cognition and behavior.

Here's a pretty insightful preprint from the Pachter lab : https://lnkd.in/gUziwec3 which showcases how both Seurat and Scanpy give different results between each other as well as between versions. I think more studies like this are necessary to bring to fruition how as computational biologists/data scientists, we should always be wary of any technical biases in our analysis pipelines, especially in scRNA-seq analysis. Below is one example, wherein calculating residual variances in scanpy is biased towards higher expression genes, while using the variance stabilization method (v2) in Seurat effectively corrects for this bias.

A new study from researchers at Flinders University aims to guide investigators on which area of the plastic bag may be most informative when sampling for DNA. Read the full article here to learn more: https://lnkd.in/e2Qtt9bC #ForensicScience #DNAAnalysis #CriminalForensics #ForensicDNA

Continuous improvement… If your segmentation is poor your analysis is poor!

Spatial Single Cell data starts with accurate cell segmentation. Full stop. The post below highlights a new tool available to CosMx SMI users to ensure transcript assignments are accurate in a spatial setting. Spatial transcript location accuracy combined with the highest plex panels available makes for a powerful combination. ????????

New group paper: Human iPSC-Derived Neurons with Reliable Synapses and Large Presynaptic Action Potentials https://lnkd.in/eniRZtPH

Weakness of AlphaFold3: I discussed Alohafold3 with Gemini (a Google-based AI chatbot) and had an interesting conversation. She understood the method well, although the code is not publicized yet, and explained its weaknesses effectively. Here’s the summary notes after a few discussions: (1) General weaknesses as an AI-based system (2) AF3 specific weaknesses (3) Hallucinations —————————————————————— (1)AlphaFold3, like most AI models, definitely shares some of the general weaknesses of AI: - Data Dependence: AlphaFold3 relies on the quality and quantity of data it's trained on. If the training data has biases or limitations, the model's predictions might reflect those. For instance, if there's a lack of data on certain protein structures, AlphaFold3 might be less confident in its predictions for those types of proteins. - Limited Explainability: While AlphaFold3 provides a confidence score, it can be challenging to understand exactly how the model arrives at its final structure prediction. This lack of transparency can make it difficult to pinpoint potential errors or areas for improvement. —————————————————————— (2) weaknesses specific to AlphaFold3 Accuracy Limitations: While AlphaFold3 is incredibly powerful, its predictions aren't always perfect. Experimental validation is still crucial to confirm the accuracy of its predictions, especially for complex proteins or those with unusual structures. Challenges with Interactions: While AlphaFold3 has improved its ability to predict protein-protein interactions, it's still an ongoing area of development. The model might struggle with particularly complex or transient interactions. Limited Data on Conformational Changes: The training data for AlphaFold3 likely focuses on protein structures in their "resting" state. Data on the full range of conformations a protein might adopt under various conditions might be limited. This can make it difficult for AlphaFold3 to predict how a protein might change shape in response to specific stimuli. —————————————————————— (3) Hallucinations Here are a couple of potential examples of hallucinations (inaccuracies) in AlphaFold3's protein structure predictions: Incorrect Folds for Proteins with Unseen Folds: AlphaFold3 is trained on a massive dataset of protein structures, but there will always be some proteins with unique folds not present in the training data. In these cases, AlphaFold3 might predict a structure that is physically impossible or highly unlikely for that protein, essentially hallucinating a fold it hasn't encountered before. Overconfident Predictions for Low-Quality Data: AlphaFold3 outputs a confidence score alongside its structure prediction. However, the model might be overly confident in its predictions, especially when trained on limited or noisy data about a specific protein. This could lead researchers to believe a potentially inaccurate structure is correct. —————————————————————— (4) someone else’s example (reposting)

I've been preaching for a couple of years that Spatial Transcriptomics Segmentation errors are the dominant biased error term in this field, not FDR (that statistic doesn't even have spatial coordinates in the equation). Check out our brand new and enabling image segmentation "transcript purifier", that significantly eliminates transcripts ending up the WRONG cell. Amazing work from two of my brightest scientists! https://lnkd.in/g2RuMYDG

A reminder that once again, segmentation is important… and the smart people at Bruker Spatial Biology have worked on tools to make transcript assignment better. Great work Joe Beechem and team! Check it out Shamini Ayyadhury ! #spatialbiology #noperfectsegmentation #spatialtranscriptomics #dataanalysis