Multi contig implementation

[JFsanchezherrero]

Dear developers,

I forked the master version of IslandPath-DIMOB (commit: 34aad6c), then merge the gff branch commit (c24711b) and add additional features.

I implemented multi-contig analysis, following your criteria and using the same functions. I only took into account that the dinuc island and dimob island generated was within the same contig.

In order to implement I had to do some modifications:

Change :

• Parsing of the ptt table and always return as key the start_end coordinates and the sequence id.

Fix some bugs:

• Fix smartmacth experimental warning message
• Fix dinuc bias loop iteration bug genomicislands cal_dinuc bug #8 (comment)

Change input:
I increased the input that were set as fixed variables in the code.
For example, users can supply the minimun number of genes under a dinuc bias (Default 8) or the minimun GI size (Defaulta 8000).

Change output:

• Apart from the GFF output, I think it might be useful to provide the dinuc bias analysis so a csv file is provided.
• Also, for each dimob genomic island identified I provide the annotation and details of the proteins involved. I guess I might have solved the issue 4 (Add generation of fasta and gbk for each GI as standard result files #4)
• Also, I provide a list of discarded regions due to length restrictions.

Add example:

• I have added again the examples files you previously discarded in order to have some sets to rely on. I could always retrieved the same results using the version and the original code.

• I tested the multi-contig feature behaviour with some assembly drafts I have here in genbank format and it worked for me. I can not deliver the data due to confidential issues but it might be appropiate to include some available on NCBI genomes for testing purposes. I downloaded and tested a genome of S. aureus (E.g. https://www.ncbi.nlm.nih.gov/assembly/GCA_900457655.1) containing several contig/scaffolds in genbank format (gbf) that I have also included in the example folder.

I hope you find useful this new implementation. If you think it might be appropiate to pull the request do it, if not, that is fine. I would be using this new implementation within my own pipeline, citing you accordingly.

Thank you very much
Please contact me if any further details are necessary