Ignoring weird tRNA from Aragorn

[JeanMainguy]

In rare cases, Aragorn predicts RNA genes with negative coordinates.

For example, in contig NZ_CSQN01000072, it predicts the following tRNA:

>NZ_CSQN01000072.1
1 gene found
1   tRNA-Ile                      c[-3,86]	34  	(gat)

Previously, start and stop positions were not validated, allowing such erroneous coordinates to persist. When written to the pangenome file, which stores coordinates as positive integers, the -3 was converted to 4294967293.

Now, with the updates made in the development branch (particularly in PR #206, which includes management of joined coordinates), extensive checks are performed on coordinates.

It has been decided that PPanGGOLiN should ignore genes predicted by Aragorn with invalid coordinates. tRNAs are only predicted to remove any CDS that would overlap them and have no other use in the PPanGGOLiN.

This PR addresses the issue in two scenarios:

1. If Aragorn predicts a gene with negative coordinates, this gene is ignored and not written to the pangenome file.
2. If a user uses a pangenome file created with version 2.0.5 or earlier that contains invalid tRNA annotations, these genes are ignored when reading the annotations to maintain compatibility between versions.

[JeanMainguy]

Here, is the contig file to reproduce Aragorn gene prediction with negative coordinates with the following command:

aragorn -t -gcbact -l -w NZ_CSQN01000072.fna.txt

NZ_CSQN01000072.fna.txt