[relipmoc]

Thanks for the good news! Is the option "-i, --intelligent For mate-pair mode, whether to redistribute reads based on junction information; (no)" ? So just keep it as default will avoid getting longer reads right?

Yes, you are right!

BTW: I found that most of the users just use skewer for preprocessing Long Mate Pair reads. It seems that few people noticed the competence of skewer to trim adapters from small RNA sequencing data.

[marghi]

Hello.

First of all thank you for skewer, it's really amazingly fast!

I have some problems in trying to achieve a certain trimming behaviour with skewer, and I am wondering if I am missing something. I am aware that there are plenty of alternative ways to get what I want, but I'd like to know the correct skewer answer nevertheless.

I have 51bp Illumina reads barcoded with a3bp barcodes located at the 5' end of each read and I would like to use skewer to demultiplex the data into the single experiments. I did try the following command line (e.g. for barcode GAT):

skewer -m head -x GAT -r 0 -d 0 -k 3 small_test.fastq -o GAT_small_test

Which works fine as long as there is a GAT at the 5' end of the read. What I wasn't expecting is that in case there's no GAT at the beginning of the read but there is one inside the sequence, then the trimming is done. For example:

before trimming: ACTCAGCNGGAAAACCTCGCCCAGATTCAGGCGTGTAGTATGCCGTCTTCT
trimmed: TCAGGCGTGTAGTATGCCGTCTTCT

Is there any way to avoid this and really restrict the trimming to the 5'end?

[relipmoc]

Hi marghi,
Thank you for your feedback! This use case was not considered by skewer. We'll add the codes for processing it in the future.

Hello.

First of all thank you for skewer, it's really amazingly fast!

I have some problems in trying to achieve a certain trimming behaviour with skewer, and I am wondering if I am missing something. I am aware that there are plenty of alternative ways to get what I want, but I'd like to know the correct skewer answer nevertheless.

I have 51bp Illumina reads barcoded with a3bp barcodes located at the 5' end of each read and I would like to use skewer to demultiplex the data into the single experiments. I did try the following command line (e.g. for barcode GAT):

skewer -m head -x GAT -r 0 -d 0 -k 3 small_test.fastq -o GAT_small_test

Which works fine as long as there is a GAT at the 5' end of the read. What I wasn't expecting is that in case there's no GAT at the beginning of the read but there is one inside the sequence, then the trimming is done. For example:

before trimming: ACTCAGCNGGAAAACCTCGCCCAGATTCAGGCGTGTAGTATGCCGTCTTCT
trimmed: TCAGGCGTGTAGTATGCCGTCTTCT

Is there any way to avoid this and really restrict the trimming to the 5'end?

[marghi]

Thank you very much!

[relipmoc]

Hi marghi,
Please download version 0.1.127 and run the following command:

$ skewer -m ap --barcode -x GAT -r 0 small_test.fastq -o GAT_small_test

You will get what you want.

[marghi]

Thank you so much! I will try right away.

Best regards

[marghi]

Dear Replimoc,

My apologies for coming up with this with so much delay, but I tried the command you recommended after upgrading to 0.1.127 and I get an error (segmentation fault). Do you have any idea of why this is happening?

Best regards

[relipmoc]

Dear Replimoc,

My apologies for coming up with this with so much delay, but I tried the command you recommended after upgrading to 0.1.127 and I get an error (segmentation fault). Do you have any idea of why this is happening?

Best regards

Sorry for the inconvenience brought to you! Could you please send an email to me about the command you used as well as a minimum dataset that can cause the segmentation fault?

[marghi]

The command I used is simply the one you posted few messages ago, i.e.:
skewer -m ap --barcode -x GAT -r 0 small_test.fastq -o GAT_small_test

Any of the test sets I tried gives the error, so I assume it's the command itself. If this is not the case I can gladly share a sample set (to which email?).

Regards,
M.

[relipmoc]

It runs well in my server. Could you please send the small_test.fastq to xxx@xxxxxx (see the skewer paper) Thank you!

BTW: are you sure that you used version 0.1.127? This version fixed some bugs in previous versions that can cause segmentation fault.

The command I used is simply the one you posted few messages ago, i.e.:
skewer -m ap --barcode -x GAT -r 0 small_test.fastq -o GAT_small_test

Any of the test sets I tried gives the error, so I assume it's the command itself. If this is not the case I can gladly share a sample set (to which email?).

Regards,
M.

[marghi]

Hi Replimoc,

Yes, I am sure I am using version 0.1.127 (Last update: August 5, 2015). I am sending via email the test set on which I get the seg fault and what causes it.

Thank you once again for your prompt support!

[dkainer]

Can Skewer do the requested operation (i.e. only taking a barcode from the start of the read) in paired end mode (PE)? or only in amplicon mode (AP)?

[lunare]

Reads are longer than expected after using skewer with -i

Hello,

I recently started using skewer. Could you provide more information on what the -i option does? The documentation only says that it will "intelligently redistribute" the reads according to junction adapter information.

I am working with Nextera mate pair data, and I used the following command to remove the adapters:

./skewer-0.1.127-linux-x86_64 -m mp -i ~/1_1_1_3kb_R1.fastq ~/1_1_1_3kb_R2.fastq

After this a subset of my reads are longer than the initial size, I don't understand how that could be. I ran it without the -i and I don't get that problem.

Some clarification would be greatly appreciated.

-----I apologize I found that this was already answered----- I think the documentation should explain this more clearly.

[dagarfield]

The -m any option

Hello!

A quick question about your "-m any" option.
In rare instances, the Nextera kits can give you sequences that look a bit as follows:

ATTAAAAATTAAAAAGAAAAGGATTATAACCTTTATAAATGGGGTATGAACCCAGTAGCTTAATTAGCTTATCTTCTGTCTCTTATACACATCTGACGCCTGTCTCTTATACACATCTCCGAGCC

The adaptor sequence we're looking to trim is 'CTGTCTCTTATACACATCT'. Generally, this sequence occurs once, and the the '-m tail' (or PE option, I think) works fine. In this (rare) case, annoyingly, the Tn5 seems to have inserted twice here. Using the '-m any' option seems to do the trick -- both instances are remove along with all the 3' sequence. But what if the adaptor is found closer to the 5' end? Will the sequence trimmed always be 3'?

Cheers,

DG

[rna_dna]

Hello relipmoc

I am hoping someone is still monitoring this thread--it looks like no one has posted here in about 4 months.

I am having trouble using skewer and I could use some help. I will wait to see if this thread is still active before going into the issue.

Thanks.