STAR alignment

[dade7]

Hi,
are there any recommendations on the STAR alignment parameters best to use before running CICERO?
Thanks for your help!

[adthrasher]

We don't have official recommendations for STAR parameters. The St. Jude Cloud team describes an RNA-Seq alignment workflow (including STAR parameters) here: https://stjudecloud.github.io/rfcs/0001-rnaseq-workflow-v2.0.0.html.

STAR --readFilesIn $ALL_FASTQ_R1 $ALL_FASTQ_READ2 \ # FastQ files, separated by comma if there are multiple. The order of your R1 and R2 files has to match!
     --outSAMattrRGline $ALL_RG \                   # Read group lines in the same order as `readFilesIn` (derived from earlier `samtools split` step).
     --genomeDir $STARDB \                          # Directory containing the STAR genome
     --runThreadN $NCPU \                           # Number of threads to use. You must request the correct amount from HPCF first!
     --outSAMunmapped Within \                      # Keep unmapped reads in the final BAM file.
     --outSAMstrandField intronMotif \              # Preserve compatibility with Cufflinks by including the XS attribute (strand derived from intron motif).
     --outSAMattributes NH HI AS nM NM MD XS \      # Recommended SAM attributes to include for compatibility. Refer to manual for specifics.
     --outFilterMultimapScoreRange 1 \              # Ensures that all multi-mapped reads will need to share the mapping score.
     --outFilterMultimapNmax 20 \                   # Max number of multi-mapped SAM entries for each read.
     --outFilterMismatchNmax 10 \                   # Max number of mismatches allowed from alignment.
     --alignIntronMax 500000 \                      # Max intron size considered.
     --alignMatesGapMax 1000000 \                   # Max gap allowed between two mates.
     --sjdbScore 2 \                                # Additional weight given to alignments that cross database junctions when scoring alignments.
     --alignSJDBoverhangMin 1 \                     # Minimum overhang required on each side of a splice junction. Here, we require only one base on each side.
     --outFilterMatchNminOverLread 0.66 \           # 66% of the read must be perfectly matched to the reference sequence.
     --outFilterScoreMinOverLread 0.66 \            # Score must be greater than 66% of the read length. So for RL=100, the alignment must have a score > 66.
     --limitBAMsortRAM $RAM_LIMIT \                 # Amount of RAM to use for sorting. Recommended value is [Max amount of RAM] - 5GB.
     --outFileNamePrefix $OUT_FILE_PREFIX \         # All output files will have this path prepended.
     --twopassMode Basic                            # Use STAR two-pass mapping technique (refer to manual).

[dade7]

Thanks.