search

CK7 / abawaca

abawaca is a binning program for metagenomics
13 stars 2 forks source link

Usage and Maximum number of dimesions #1

Open fhcoutinho opened 9 years ago

fhcoutinho commented 9 years ago

Hello, could you please provide a more detailed information on the input files? Is there a limit to the number of dimensions that can be used? I have been trying to run ABAWACA with .lrn and .names file formatted exactly like the ESOM input. I am using Kmer content, abundance and coverage of contigs across samples and also taxonomic affiliation (In a total of 3471 data points and ~14700 dimensions). However, I keep getting errors like these if I input a .lrn file with more than 1510 dimensions:

src/ClusterData.cpp (135): Fatal error: number of dimensions found is 1, expected to find 1510: 0991055075714494

Is this some kind of memory issue that causes the input files not to be read correctly? I already checked the files in several ways to make sure all dimensions are actually there, with numeric values, and unique names.

Any help is much appreciated.

CK7 commented 9 years ago

Hello fhcoutinho, My apologies for the very late response. This version of abawaca is an old one and to be honest I do not recommend that you use it because it takes a lot of time to finish. I published the code as part of our publication (Brown et al., Nature, 2015) so that people can try it in case they want. If you still want to try this version then I'll be happy to help with your issue - please send me the .lrn and .names files and I will take a look. I am currently working on a newer version that is going to be much faster (and provide better results) so stay tuned!

Cheers Itai

fhcoutinho commented 9 years ago

Thank you, I was able to run it by using .lrn and .names file in the same format used for ESOM. Please do let me know when the new version is released.

Cheers,

Felipe Hernandes Coutinho, MSc.PhD student,Laboratório de MicrobiologiaInstituto de Biologia/Universidade Federal do Rio de JaneiroIlha do Fundão - Zip code: 21944-970 - Rio de Janeiro, RJ - Brasil - P.O. Box: 68011Phone: (21) 39326567

Date: Wed, 30 Sep 2015 12:42:29 -0700 From: notifications@github.com To: abawaca@noreply.github.com CC: fhcoutinho@live.com Subject: Re: [abawaca] Usage and Maximum number of dimesions (#1)

Hello fhcoutinho,

My apologies for the very late response.

This version of abawaca is an old one and to be honest I do not recommend that you use it because it takes a lot of time to finish. I published the code as part of our publication (Brown et al., Nature, 2015) so that people can try it in case they want. If you still want to try this version then I'll be happy to help with your issue - please send me the .lrn and .names files and I will take a look.

I am currently working on a newer version that is going to be much faster (and provide better results) so stay tuned!

Cheers

Itai

— Reply to this email directly or view it on GitHub.