% unmapped reads

bioinfosm · 11-23-2009, 10:34 AM

Hi all,

For any solexa run, we see 60-80% passing filter reads, of which 70-90% map to the reference sequence. There are quite a few reads that map to adaptors and other random sequences.

Does anyone know of a resource to get an average for those passing-filter reads? that map to adapter, etc and are essentially noise and not useful. I remember it being mentioned in some review as well, but cannot recollect

thanks..

der_eiskern · 12-11-2009, 04:05 PM

i'm not sure what exactly you want ("an average for those passing-filter reads")

but i've also noticed often two-thirds of the 30-10% of unmappable reads are alignable if you can tolerate a higher error rate.

from cloning random fragments i've noticed that 1/10-1/100 of cloned fragments from a sequencing library will have a truncated adaptor or multiple adaptors. since i didn't get my adaptors from illumina directly this could reflect contamination of my modified IDT oligos. it varied between library preps using the same batch of adaptors.

can you not just make a fasta file of the adaptor and iteratively truncate your read during mapping to such a fast a file to estimate it? i'd be more interested in seeing what what doesn't pass filter looks like, any idea how to get that data from the pipeline?

Xi Wang · 12-11-2009, 10:04 PM

Quote:

Originally Posted by der_eiskern

from cloning random fragments i've noticed that 1/10-1/100 of cloned fragments from a sequencing library will have a truncated adaptor or multiple adaptors. since i didn't get my adaptors from illumina directly this could reflect contamination of my modified IDT oligos. it varied between library preps using the same batch of adaptors.

I am wondering what makes the sequencer not report the adaptor sequences and exactly the very beginning of what we want to sequence. Or it will be quite normal to sequence the tails of adaptors.

Quote:

Originally Posted by der_eiskern

can you not just make a fasta file of the adaptor and iteratively truncate your read during mapping to such a fast a file to estimate it? i'd be more interested in seeing what what doesn't pass filter looks like, any idea how to get that data from the pipeline?

Those not passing the filtering usually are with low quality scores, which indicates that the base calling may be incorrect or it is hard to call bases.

aleferna · 07-04-2010, 10:15 AM

Does anybody know what is the expected percent of adapter sequence that one would expect in a Solexa sequencing run?

I'm mapping this solexa paired end run and I had to mask away about 50% of the sample because it maps to the adapters. Is this normal? If so and we do this again, how would you minimize the amount of adapter sequence that you get in the library????

If it is not normal, is there any doc/spec that we can use to complain to the sequencing service and get it done properly?

Xi Wang · 07-04-2010, 09:46 PM

Quote:

Originally Posted by aleferna

Does anybody know what is the expected percent of adapter sequence that one would expect in a Solexa sequencing run?

I'm mapping this solexa paired end run and I had to mask away about 50% of the sample because it maps to the adapters. Is this normal? If so and we do this again, how would you minimize the amount of adapter sequence that you get in the library????

If it is not normal, is there any doc/spec that we can use to complain to the sequencing service and get it done properly?

It may depend on what you sequenced. RNA-seq? miRNA-seq?

aleferna · 07-05-2010, 01:10 AM

just plain human DNA

Xi Wang · 07-05-2010, 01:17 AM

I have no experience on plain DNA sequencing. But I think your problem may largely due to sample preparation.

aleferna · 07-05-2010, 01:35 AM

But in what part, is that the main issue is that we provide the samples, but another lab did the Paired End preparation. If there was a problem in the paired end preparation we need to go back to this people and tell them that they made a mistake. Because its such an expensive test this can get really ugly and I don't have any reference to make the point...

aleferna · 07-05-2010, 01:36 AM

How much do you expect in RNA?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
unmapped reads with ssaha2	sammy07	Bioinformatics	2	10-22-2012 03:23 AM
BWA - unmapped reads	Adamo	Bioinformatics	23	02-22-2012 10:12 PM
Velvet for unmapped reads	leeht	Bioinformatics	1	11-20-2010 05:39 PM
bfast for unmapped reads	Protaeus	Bioinformatics	2	11-17-2010 03:35 PM
What are the unmapped reads	beelu	Illumina/Solexa	1	09-09-2010 06:18 AM

11-23-2009, 10:34 AM	#1
bioinfosm Senior Member Location: USA Join Date: Jan 2008 Posts: 482	% unmapped reads Hi all, For any solexa run, we see 60-80% passing filter reads, of which 70-90% map to the reference sequence. There are quite a few reads that map to adaptors and other random sequences. Does anyone know of a resource to get an average for those passing-filter reads? that map to adapter, etc and are essentially noise and not useful. I remember it being mentioned in some review as well, but cannot recollect thanks.. __________________ -- bioinfosm

07-04-2010, 10:15 AM	#4
aleferna Senior Member Location: sweden Join Date: Sep 2009 Posts: 121	Solexa Sequencing of Paired Ends Does anybody know what is the expected percent of adapter sequence that one would expect in a Solexa sequencing run? I'm mapping this solexa paired end run and I had to mask away about 50% of the sample because it maps to the adapters. Is this normal? If so and we do this again, how would you minimize the amount of adapter sequence that you get in the library???? If it is not normal, is there any doc/spec that we can use to complain to the sequencing service and get it done properly?

07-05-2010, 01:17 AM	#7
Xi Wang Senior Member Location: MDC, Berlin, Germany Join Date: Oct 2009 Posts: 317	I have no experience on plain DNA sequencing. But I think your problem may largely due to sample preparation. __________________ Xi Wang

12-11-2009, 04:05 PM	#2
der_eiskern Member Location: California Join Date: Jul 2009 Posts: 46	i'm not sure what exactly you want ("an average for those passing-filter reads") but i've also noticed often two-thirds of the 30-10% of unmappable reads are alignable if you can tolerate a higher error rate. from cloning random fragments i've noticed that 1/10-1/100 of cloned fragments from a sequencing library will have a truncated adaptor or multiple adaptors. since i didn't get my adaptors from illumina directly this could reflect contamination of my modified IDT oligos. it varied between library preps using the same batch of adaptors. can you not just make a fasta file of the adaptor and iteratively truncate your read during mapping to such a fast a file to estimate it? i'd be more interested in seeing what what doesn't pass filter looks like, any idea how to get that data from the pipeline?

07-05-2010, 01:10 AM	#6
aleferna Senior Member Location: sweden Join Date: Sep 2009 Posts: 121	just plain human DNA

07-05-2010, 01:35 AM	#8
aleferna Senior Member Location: sweden Join Date: Sep 2009 Posts: 121	But in what part, is that the main issue is that we provide the samples, but another lab did the Paired End preparation. If there was a problem in the paired end preparation we need to go back to this people and tell them that they made a mistake. Because its such an expensive test this can get really ugly and I don't have any reference to make the point...

07-05-2010, 01:36 AM	#9
aleferna Senior Member Location: sweden Join Date: Sep 2009 Posts: 121	How much do you expect in RNA?