Simplify installation and user interface

Major enhancements

1. Installation and environment management is simplified. The user is now requested to install FilTar within its own standalone conda environment

2. Modify how users specify dataset metadata. Previously users had to specify dataset metadata within nested YAML data structures. This is now simplified by requesting that users add this information in TSV format. This information is then processed by snakemake using the pandas module.

3. Simplify dependency parameter configuration. FilTar utlilises a large number of dependency applications. These can now be more easily configured by the user within the config/dependencies.yaml file

4. Use higher confidence miRNA annotations: Allow the user to use high-confidence miRNA annotations only, from miRBase. This option is set as default

Minor enhancements

1. Update README to include URLs to the FilTar publications within Bioinformatics
2. Throw an exception if the user specifies an invalid miRNA target prediction algorithm
3. Use APAtrap binary files instead of script which reduces the number of dependencies to be installed by the user

Allow configuration of genome index type

Enhancement

In this release, we enable users to configure the type of genome indexing they used to build genomes with HISAT2. In particular users have the choice of using standard (default) and splice-aware values for the genome_index_type configurable option. Using the standard genome index option instead of the splice-aware option reduces the memory requirement for this particular process from ~200GB, to ~8GB according to HISAT2's documentation.

Fix Ensembl URLs

In this release, we fix the URLs used to download data from Ensembl.org through FTP, due to changes on the ensembl server itself.

In particular, the previous /ensembl/pub/ pattern is substituted with the following pattern /pub/ in all ensembl URLs used in the project.

Fix issues with calling APAtrap perl scripts

Bug fixes

1. Explicitly call APAtrap perl scripts using the version of perl in the environment, rather than the version of perl referenced in the shebang line of the perl scripts themselves
2. Fix miRanda wildcard recursion issue
3. Add missing 'temp' tags within the 'get_utr_and_cds' module Snakefiles
4. Alter the species config files to in order to remove chromosomes without annotated 3'UTR sequences, which leads to errors

Simplify usage by removing option data from output file name

Enhancements

1. Greatly simplify usage by deleting most intermediate files automatically rather than relying on the user to do this
2. Make miRNA target prediction algorithm an option, rather than have the user specify this in the output file name
3. Further simplification of the final output target file name ("target_predictions.txt")
4. Allows the user to use both fastq.gz and fq.gz file extensions for RNA-Seq data
5. Enable processing when no RNA-Seq data, and hence no expression filtering is applied
6. Add a header to miRanda output

Bug-fixes

1. Enable use of TPM thresholds when not reannotating 3'UTRs
2. Allows transcript ID with version numbers in the double digits
3. Fix a big with the automated TargetScan test
4. Fix bugs associated with empty files, if for example, no 3'UTRs are extended

Automate download of APAtrap dependency

Enhancements

1. Automate the download of APAtrap scripts from source forge (https://sourceforge.net/projects/apatrap/files/Source%20Codes/)

Bug fixes

1. Resolve ambiguous target file names for some miRanda rules, which was leading to a recursion error

Initial Release

The version of FilTar described in this preprint: https://doi.org/10.1101/595322