Hello! I’m interested in running some loci through the L2G models that have been produced by OpenTargets, and have been exploring the gentropy package and documentation to do this.
I’d be really interested to know if there is any example code/write-ups showing best practices for running novel loci through the “locus_to_gene_feature_matrix” steps, and subsequently predict with a pre-trained model with the “locus_to_gene” step, please?
I’ve read through this topic by nabuan that points to the gentropy package as the best place to start. There’s also mention that the data required to run the pipeline is planned to be released at some point or could be generated from scratch to fit into the pipeline.
Any advice on how to best approach this would be much appreciated, thanks! 
Hi James,
Thank you for your question and your interest in our gentropy package!
As our March release is soon approaching, we suggest that you wait for the datasets required for the pipeline to be made public. This will allow you to run the pipeline with your new loci.
Further guidelines/example code will be made available soon after 
Best wishes,
Vivien
2 Likes
Thanks for the quick reply @vivienho - I’ll keep an eye out for more information in the new release!
Best wishes,
James
Hi again! I’ve been doing some searching but couldn’t find a date for the next release - is there a specific day scheduled?
Hi @JAnomics,
The release is currently planned for next week, and I will post in the Community release channel when it is live 
Thanks,
Helena
1 Like
Hi @hcornu ,
Excited to see the new release! I’m still having trouble finding information about running the L2G models on new data. For example, how to setup/generate a FeatureMatrix for use in a pre-trained model, and knowing what data I need to have locally to do that? Any documentation or guidance would be much appreciated!
Best wishes,
James