Hi everyone,
My name is Keaun Amani and I'm the Founder of Neurosnap.ai, a platform for making access to bioinformatics as simple as possible.
As I'm sure you're all aware, accessing, deploying, and maintaining bioinformatic models and tools can be quite a pain at times. Neurosnap aims to alleviate these pain points by providing a straightforward graphical web interface where you can directly access all your favorite tools and models like AlphaFold2, DiffDock, RoseTTAFold All-Atom, RFdiffusion, and many more without any hassle.
Simply select the tool you'd like to access, configure it, upload your desired inputs, and go grab a coffee while everything runs on our servers. Once the results are complete you'll get an email and be able to access the results as well as intuitive visualizations directly on our website.
As of right now our platform supports the following task types:
- Drug Design
- Inverse Folding
- Molecular Docking
- Molecular Dynamics
- Protein Annotation
- Protein Clustering
- Protein Conformations
- Protein Design
- Protein Expression
- Protein Folding
- Protein Localization
- Protein Solubility
- Protein Stability
- RNASeq / Transcriptome Analysis
- Signal Peptide Detection
- Toxicity Prediction
We also have some fun demonstration videos on youtube: https://youtu.be/L4gW-sdVdWY
Anyways, I hope our platform provides value to your research endeavors and if you have any questions please feel free to reach out to us!
For RNA-seq analysis, it asks me to upload an interleaved compressed FASTQ through my browser.
Are you kidding me?
Haha, yeah though we also support compressed FASTQ files and for even bigger uploads you can contact support and we'll have our servers download the file directly from one of your own servers.
Hey Keaun,
Thanks for sharing. How does your product compare to other web portals for running ML models for biology, like tamarind.bio?
Great question Kevin! The main differences are in ease of use and accuracy. Many of our users have reported that they have tried both but prefer using our platform due to improved accuracy and preference for look and feel. After all, we have also been operating for longer as well.
For accuracy, what really sets us apart is that we have developed additional metrics & optimizations on top of many of our services that make them more reliable and intuitive to use. For example, our AlphaFold2 is built on top of the ColabFold tool, which has been found to improve structure prediction quality due to improved MSA generation.
We take this a step further by building on top of the mmseqs2 MSAs generated from ColabFold, along with some additional sampling steps, which we have found to substantially improve accuracy for certain proteins. For AlphaFold2, we also generate various metrics that Tamarind and other platforms do not offer which gives our users a greater ability to discern between an inaccuracies.
Ultimately, we encourage people to use whatever works best for them (but neurosnap is better ;) ).