GoekeLab / m6anet

Detection of m6A from direct RNA-Seq data
https://m6anet.readthedocs.io/
MIT License
100 stars 17 forks source link

Questions about the m6anet inference results #133

Open mumdark opened 9 months ago

mumdark commented 9 months ago

Dear Developer,

Thank you for developing such an great tool !

I have a question when running m6anet inference.

Why are there only a few positions, specifically 16, retained in the output? In my data, there are far more positions with more than 20 reads.


Here are the read counts for my input data(a total of 1,117 positions with more than 20 reads):

image

Here are my results:

image

Here's my code:

m6anet inference --input_dir m6anet_datapre/ --out_dir m6anet_inference/ --n_processes 4 --num_iterations 1000

Thank you for your response! Much appreciated.

mumdark

mumdark commented 9 months ago

It appears there was an issue when running m6anet dataprep for data preparation, which might have resulted in the unexpected final output.

image

How to solve this problem? Thanks.

JSluo888 commented 6 months ago

I am having the same issue

soniacruciani commented 1 month ago

Hello, I also am encountering this problem. I checked my .json file and it is fine, I have 14 positions which I know should be picked up (and were on a different fast5 file of the same exact experiment), but in my output I only have 13. I have ~2k reads coverage per position, and ~800 reads at least for each position in dataprep output.

Any idea of what might be happening?

Thanks!