brendanofallon / jovian Goto Github PK

Hi,

I have some HG002 long reads that I have error-corrected to almost short-read quality - https://gist.github.com/jelber2/02fdca09ed3b12b650b512ba200f3b67

I was wondering if there might be a tool to shred my alignments say from 20000-bp (just for example) to 150-bp or whatever the maximum read length allowed by jenever is. I guess I could try to make either a Rust (or Julia) tool to do this, but I was asking just in case someone has something already. I know of some tools for unaligned reads, but not for already-aligned reads.

Best,
Jean

Dear Brendan,

I know that you were considering something else in terms of making a long-read model, but I wanted to try and make a model and classifier based on these CRAM files, after shredding to 150-bp alignments pieces

, but I am a little confused as

seems to suggest that the fourth column of the bed file for the config.yaml (from GIAB for example) would be subdivided into the value classes one sees below

Am I correct that such a bed file would be subdivided based on the vals_per_class? I would really be interested in seeing how well a long-read model (even a "shredded" one) for jenever performs on Nanopore SUP_Herro data that I had sent you links for.

Should this read python3 and not python ? Had to make a symbolic link in my container for Ubuntu 20.04 pointing from python3 to python.

https://github.com/brendanofallon/jovian/blob/dcbd3c5e133f59e3cefd9ea0212132109896bfea/setup.py#L28C1-L28C18