search

marbl / ModDotPlot

MIT License
101 stars 7 forks source link

Some feedbacks #1

Closed oushujun closed 1 year ago

oushujun commented 1 year ago

Hi Alex,

Great tool! Congratulations!! While using it, I found several places that could benefit from more clarification and slight modifications.

  1. The relationship between -w and -d is not very clear.
  2. In the bed file, query_name and reference_name are all replaced by the -p value. Can it be the original sequence names?
  3. Line 34 of mod_identity.py is hard coded to perID*100 >= 80; can you make it customizable to allow reporting lower identity regions?
  4. The mod_identity.py script can only run on one sequence? Can it take the entire genome?
  5. The plot.r script requires a folder named results in the working directory; otherwise, it will run into errors.
  6. The plot.r script requires ggplot2, dplyr, scales, RColorBrewer, data.table, cowplot, glue, argparse packages installed in R.

Thanks! Shujun

alexsweeten commented 1 year ago

Hi Shujun. Thanks for the feedback, and I'm so sorry for the (super) late reply!

The relationship between -w and -d is not very clear.

I agree. I have renamed -w to -r, referring to the resolution of the dotplot. Essentially this represents how many chunks to divide the sequence up into for comparison.

In the bed file, query_name and reference_name are all replaced by the -p value. Can it be the original sequence names?

Fixed in afc0344e3669b50df30f48560dbbccc689012236

Line 34 of mod_identity.py is hard coded to perID*100 >= 80; can you make it customizable to allow reporting lower identity regions?

Fixed in afc0344e3669b50df30f48560dbbccc689012236. This is now adjustable via the -id command.

The mod_identity.py script can only run on one sequence? Can it take the entire genome?

I have refactored parse_fasta to utilize pysam. This will generate a bed file for each header sequence in the input fasta.

The plot.r script requires a folder named results in the working directory; otherwise, it will run into errors.

I plan to substitute the Rscript with an alternative in Python. This will reduce the number of dependencies required, and also unify with the interactive component of Mod.Plot!

Thanks again! I hope all is well in Columbus :)

oushujun commented 1 year ago

Thank you, Alex! All is well here :)