Quality score and loss of reads

[laura benestan]

Hi everyone,

I received my RAD sequencing data last month, which were pretty good, with an amount of 185 millions read per lane, on average. 

However, when after the process rad tag step (10 quality score at minimum, 0.15 window sliding), just 60 to 88 percent of all the reads were kept for future analyses. 

The quality scores of my libraries are quiet low, ranging from 30 to 35, nevertheless, is it normal to keep such a little amount of data ?

Is there anybody who has already experienced this kind of results ?

Thanks,

Laura

[Ryan Waples]

Quality scores of 30 - 35 on the phred scale are good and about as high as you can hope for.  (30 = estimated 99.9% base call accuracy ), see http://en.wikipedia.org/wiki/Phred_quality_score.

But Perhaps I am misunderstanding you.

That percentage of retained reads sounds about right to me.

Perhaps others will weigh in.

Ryan

--
--
For more options or to unsubscribe: http://groups.google.com/group/stacks-users
Stacks website: http://creskolab.uoregon.edu/stacks/
 
---
You received this message because you are subscribed to the Google Groups "Stacks" group.
To unsubscribe from this group and stop receiving emails from it, send an email to stacks-users...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[bbarker505]

Laura,
Which sequencing platform did you use? We have had issues with getting back low quality R2 reads from Illumina HiSeq. After talking with Illumina tech support it sounds like this has to do with cluster re-synthesis of low diversity libraries, but I'd like to hear other researchers' experience with this.
Brit

[Laura Benestan]

Hi Brit,

Individual RAD libraries were labeled with unique 6 bp barcodes. Per lane, 48 individuals were pooled in equimolar proportions and sequenced
on an Illumina HiSeq2000. We used the sequencing platform of MacGill University at Montreal (Canada).
Besides, I have read several articles which indicate on average about 70% to 90% of read retained after the quality score filter.
Therefore, I think my results are not surprising.
In my lab, some students obtained more than 90% (of retained reads)
However, all of them have one thing in common : they use restriction enzyme with few base pair.
Indeed, much more you get short fragment ofr sequencing, much more the amplification is efficient and I presume it may have less error...