add --outMultimapperOrder Random --outSAMmultNmax 1 to aligner options for STAR in mRNAseq workflow

[katsikora]

to assure feature counts doesn't inflate multi-mapping alignments estimate and that the estimate of fragments counted on exons is interpretable

also raised here maxplanck-ie/snakepipes#1165