number of mismatches

[Olivier SAULNIER]

Hi all,

I moved to STAR v2.5.0 for PE RNA-seq data 2x101.
I have a really good % of  uniquely mapped reads.
But when i'm looking on IGV, i can see some reads with a really bad alignement (cigar 37M64S : btw i don't know why missmatches are called S ?).
How this read can be aligned, even if it mate is well aligned (cigar 101M) ?

COMMAND="${STAR} --runThreadN 8 --runMode alignReads --genomeDir ${ID} --readFilesIn ${DATA}/${SAMPLE}_1.fastq ${DATA}/${SAMPLE}_2.fastq --sjdbGTFfile /data/annotations/Galaxy/RefGene/hg19/hg19_refGene.gtf --outSAMattributes All --outFileNamePrefix ${JOBNAME} --outFilterMultimapNmax 1 --outFilterMismatchNmax 3 --outSJfilterCountUniqueMin 3 3 3 3 --outSJfilterCountTotalMin 3 3 3 3 -sjdbOverhang 100 --outSAMmultNmax 1 --outSAMtype BAM SortedByCoordinate  --outReadsUnmapped Fastx --outBAMsortingThreadN 8"

Many thanks,
Olivier S

[Alexander Dobin]

Hi Olivier,

S in the CIGAR represents "soft-clipping", i.e. bases that could not be aligned and were trimmed from the ends of the reads.

The mismatches inside the soft-clipped ends are not counted by STAR (even though IGV will show them as highly mismatched tails). If you want to control the amount of soft-clipping, please look at this post:

https://groups.google.com/d/msg/rna-star/UCUM5x9cXGk/VMuS0ZZtJMEJ

There are many reasons why the 64nt from this particular read could not be mapped: (i) poor sequencing quality tail; (ii)A-tail; (iii) Adapter; (iv)hard-to-detect junction; (v) chimeric/fusion junction.

To detect the (v), you can switch on chimeric detection and see whether this read will be reported as chimeric. You can also BLAST the sequence.

Cheers

Alex