Problems with STAR alignment

tellsparck · 05-25-2015, 07:38 PM

Hi,
I hope you can help me with a problem we ran into while aligning reads from a 100bp PE library with average insert size of 200bp. About 74% of the reads fall into the category "% of reads unmapped: too short". I also noticed that the average mapped length for this library is 184.4. Other libraries sequenced alongside had average mapped length little bit above 190 (still shorter than 200, but better) and their uniquely aligned percent is much better (75-80%).I do expect significant rRNA contamination because of the library prep method used. Given that the other libraries prepped and sequenced (in the same lane)along with the failed one performed better, could this be an issue with smaller insert size? Or is it a sequencing quality issue? Is there a way to "rescue" the failed library?
Thanks for your help!

jwfoley · 05-26-2015, 07:02 AM

Are you removing sequencing adapters? Maybe that library is all adapter dimers, so after they're trimmed there's nothing left.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
2nd Pass Alignment with STAR for RNA-sequencing	cbaudo	RNA Sequencing	3	03-04-2015 04:59 AM
STAR for alignment	sbdk82	Bioinformatics	2	09-09-2014 09:22 AM
STAR: Alignment reported with less matches than parameter --outFilterMatchNmin	AnneBiton	Bioinformatics	0	07-29-2014 12:45 PM
Problems with htseq-count reading bam file created by STAR	priya	RNA Sequencing	3	06-02-2013 12:32 AM
STAR-alignment/confusion with mismatches & multi-mapping parameter	babi2305	Bioinformatics	9	05-07-2013 01:45 PM

05-25-2015, 07:38 PM	#1
tellsparck Member Location: Philadelphia Join Date: May 2012 Posts: 19	Problems with STAR alignment Hi, I hope you can help me with a problem we ran into while aligning reads from a 100bp PE library with average insert size of 200bp. About 74% of the reads fall into the category "% of reads unmapped: too short". I also noticed that the average mapped length for this library is 184.4. Other libraries sequenced alongside had average mapped length little bit above 190 (still shorter than 200, but better) and their uniquely aligned percent is much better (75-80%).I do expect significant rRNA contamination because of the library prep method used. Given that the other libraries prepped and sequenced (in the same lane)along with the failed one performed better, could this be an issue with smaller insert size? Or is it a sequencing quality issue? Is there a way to "rescue" the failed library? Thanks for your help!

05-26-2015, 07:02 AM	#2
jwfoley Senior Member Location: Stanford Join Date: Jun 2009 Posts: 181	Are you removing sequencing adapters? Maybe that library is all adapter dimers, so after they're trimmed there's nothing left.