Stpipeline

* Made the parsing of unique UMIs gene by gene and parallel
* AdjacentBi is now the default method for UMI counting
* Made the trimming function output the trimmed R2 in BAM format with the barcode and UMI
* Made the mapping function works with a BAM file as input (latest STAR release)
* Made the annotation function parallel
* Made the quality step parallel
* Improvements in speed and memory (constant memory use)
* The STAR genome loading strategy can be now set
* Added an affinity based method to cluster UMIs
* Added option to set the STAR BAM sort memory limit

* Fixed a bug that will output the matrix of counts inverted

* st_qa.py generate expression heatmap plots
* Fixed a minor bug in the computation of the saturation curves
* adjust_matrix_coordinates now does not update the coordinates by default
* adjust_matrix_coordiantes works with the latest ST Spot detector format
* small updates in the fastq merging script
* relaxed a bit the restriction checks for some parameters

* Added extra scripts:
- merge_bcl : merge BCL files based in patterns
- filter_gene_type_matrix : filter gene in output data based on Ensembl gene types
* Bumped Pysam and HTSeq

* Small update to make the PIP installation more robust

* Small update to make the PIP installation more robust