Hello. First let me just say, thank you for developing m6anet, it has been extremely helpful in producing data for RNA modifications, especially on the RNA002 kit and hopefully soon for RNA004.
I'm currently working to identify m6a modifications using data from the RNA004 chemistry. I am using the development version of m6anet that follows f5c eventalign before running the m6anet dataprep and inference with the new rna004_kmer model. I can produce high-quality site predictions (>0.9-0.999) from this workflow, but for individual sites the accuracy hits a cap of ~0.8.
I know the development version said only site level predictions have been tested, with no results on the individual currently. Is there any available data or information on the accuracy of individual site predictions/will there be any updates regarding this soon?
Thanks again for developing m6anet, I look forward to hearing back!
Hello. First let me just say, thank you for developing m6anet, it has been extremely helpful in producing data for RNA modifications, especially on the RNA002 kit and hopefully soon for RNA004.
I'm currently working to identify m6a modifications using data from the RNA004 chemistry. I am using the development version of m6anet that follows f5c eventalign before running the m6anet dataprep and inference with the new rna004_kmer model. I can produce high-quality site predictions (>0.9-0.999) from this workflow, but for individual sites the accuracy hits a cap of ~0.8.
I know the development version said only site level predictions have been tested, with no results on the individual currently. Is there any available data or information on the accuracy of individual site predictions/will there be any updates regarding this soon?
Thanks again for developing m6anet, I look forward to hearing back!