Are there Julia packages for RNA/DNA/protein structure prediction?

I am doing research about RNA and proteins. Therefore, I am looking for a method to simulate the secondary or 3D structure. I have used some methods based on other programming language constructs. But those methods are not very suitable for my projects.

I recently started learning Julia and found that Julia is a beautiful language. Based on Julia’s high efficiency, is there a suitable package for RNA or protein structural simulation? (like ViennaRNA for RNA or Alphafold for protein simulation)

Thanks! :grin:

2 Likes

This isn’t my field so I’m not familiar with all packages/features, but have you seen https://biojulia.net/?

1 Like

For other packages you can search in the registry.I saw many packages
for biological themes in general.

1 Like

Thanks for that link!
BioStructures.jl and BioSequences.jl seem relatively relevant to my project. I am trying to use them in my project.
Thank you!

1 Like

That might be something I am finding for. I will try it, thanks! :grinning:

You won’t find anything “like AlphaFold”, but “we” (@gfahurbottino, my PhD student, actually) have developed a package to predict secondary structures and other one dimensional traits, here: GitHub - Hugemiler/TintiNet.jl: Implementation of the TintiNet architecture for Julia

The article is submitted for publication, maybe the infrastructure there serves as a inspiration. I think that could be adapted to predict RNA traits, what do you think, @gfahurbottino ?

3 Likes

You might also want to take a look at BioStructures.jl, as well as BioMakie.jl for some plotting options

3 Likes

Hi, BetaQ!

As @lmiq mentioned, we recently build a 100% julia-based (Flux.jl + Transformers.jl + BioJulia) sequence-structure translator. It’s based on a general encoder-decoder architecture, where the encoding block processes a sequence and the decoder block estimates properties.

In principle, you could use the same decoder architecture we built to train for Nucleic Acid secondary structures, rather than protein secondary structures. The training/evaluation scripts are available on the following repository (you probably want the IGBT_classifier version, for categorical secondary structures):

3 Likes