search

combogenomics / medusa

A draft genome scaffolder that uses multiple reference genomes in a graph-based approach.
http://combo.dbe.unifi.it/medusa/
GNU General Public License v3.0
42 stars 15 forks source link

w2 parameter explanation #7

Closed habibr closed 8 years ago

habibr commented 8 years ago

In the program parameters' description, it says:

The option -w2 is optional and allows for a sequence similarity based weighting scheme. Using a different weighting scheme may lead to better results.

I am a bit confused with the wordings. Which option would lead to better results? Turning -w2 on or off?

mfondi commented 8 years ago

Dear Habib, thanks for your question. Indeed, the wording in the parameter's description could be misleading. The sentence should read "Using a different weighting scheme may lead to better results than the default usage of MeDuSa" Besides the approach describe in MeDuSa's paper, we have implemented a sequence-similarity based weighting scheme that can be used to weight the edges of the scaffolding network (see additional material of the same paper). When we tested this approach on our dataset of genomes we found some improvements for some of them but not for all of them. So, you might want to try this option for your case and compare the results to the default usage of MeDuSa.

hope this helps. cheers m.

On 09/02/16 05:15, Habib R wrote:

In the program parameters' description, it says:

The option -w2 is optional and allows for a sequence similarity
based weighting scheme. Using a different weighting scheme may
lead to better results.

I am a bit confused with the wordings. Which option would lead to better results? Turning -w2 on or off

— Reply to this email directly or view it on GitHub https://github.com/combogenomics/medusa/issues/7.

Marco Fondi, PhD Dep. of Biology, University of Florence Via Madonna del Piano 6, S. Fiorentino, Florence, Italy Tel. +39 055 4574736