bug: STAR using uncompressed sample files when mate relationship is not

[balajtimate]

Describe the bug
With the revised mapping logic from #131, when the input samples are not mates and the mate relationship needs to be determined by mapping the reads against the transcripts, STAR uses the original input sample files instead of the uncompressed subset of the samples, resulting in an error, as the STAR command is currently not built to handle .gz files.

To Reproduce
Run HTSinfer in paired-end mode on files from different mates, one of them compressed, e.g.

htsinfer tests/files/first_mate.fastq tests/files/mixed_mates_compressed.fastq.gz

Expected behavior
Add the paths to the subset of samples as argument when calling to the mapping function in get_library_type.py