Ptolemy suggestions

[jlobatop]

After some months developing code for Ptolemy, there are some missing features/bugs/recommendations I would like to point out to keep track of them:

• misc/assembly/ReverseComplement.scala has been removed as it was redundant.

• Relative folders handling: if Ptolemy.jar is executed in ./ while the tab-separated file is located in ./folder/subfolder/genome_list.txt, the genomes should have the path relative to the folder from which the program is called and not with respect to the file itself.

• Given that at the beginning of AlignReads.scala, the following snippet was being used:

  val verbose = config.verbose || config.debug
  

  it was changed so that the verbose and debug flag trigger the same screen outputs. Although the alignment step has been tested and debug shouldn't be required anymore, it is here acknowledged that this behavior may be reverted.

• There are some statistical tools already implemented in /canonical_quiver/PtolemyMetrics.scala but they are not used at all, as they are not called in any submodule.

• This issue has been clarified in the README file When running the testing_data case, a warning appears (maybe due to having more strains in a node that the number of strains considered?) during the canonical-quiver step and it may require attention:

  $ java -jar ptolemy.jar canonical-quiver -s ./syntenic_anchors.txt -d ./ptolemy_db/ -o ./
  (...): Fetching hashmap Z
  (...): Fetching starting node ID
  (...): --Starting node is: 1504
  (...): Formatting syntenic anchors
  --WARNING: unexpected number of nodes in a syntenic anchor. Discarding: 994,1497,496,997
  (...): Constructing canonical quiver
  (...): Constructed canonical quiver with 521 nodes and 538 edges
  (...): Writing GFA file to disk
  (...): Successfully completed! 
  

• In the index-graph submodule there is no output folder option, having that the performed indexing of the graph is stored in the same folder as the canonical_quiver.gfa.

• Further investigation is required to check the effect of flag parameters on the results. For example, minInterSize and kmerSize have the same value by default. However, strange graph results have been obtained for runs if their value was different.

• Add a new genome to an already Ptolemy-created graph without the burden to start from the very first step (although not sure how much will this affect the current implementation).

• From a user point of view, an option to go from the genome tab-separated file to the graph may be useful in a single command, avoiding three different steps.