Undetected junction adaptor in Category D

[MesutOezil]

Hello. I have been using NextClip for these a few month and really like this program.

Let me ask a new question, the Category D. I ran NextClip on HiSeq reads (PE171) from our Nextera Mate Pair libs, and found at the end of some read in Category D, several bases probably consecutively matching the beginning of the junction adaptor. If they are genuine mates, I want to rescue and use them in scaffolding. If I rerun NextClip on the same data (and run later on future data), how would you suggest to change parameters? Of course, I have consulted with the manual, I want to listen to what the thoughtful developer would say.

Very best wishes from Japan,

Shigehiro

[richardmleggett]

Hi Shigehiro,

Thanks for your message.

The only option in the current version of NextClip is to reduce the sensitivity of the matching, using the --strict_match (-x) parameter. However, I personally wouldn’t play around with this too much, as you risk too many false matches.

Are there that many reads that are affected? I would expect it to be relatively small.

I could think about adding a feature for a later version to handle these kind of cases.

Thanks, Richard

On 9 Jun 2014, at 06:04, MesutOezil notifications@github.com<mailto:notifications@github.com> wrote:

Hello. I have been using NextClip for these a few month and really like this program.

Let me ask a new question, the Category D. I ran NextClip on HiSeq reads (PE171) from our Nextera Mate Pair libs, and found at the end of some read in Category D, several bases probably consecutively matching the beginning of the junction adaptor. If they are genuine mates, I want to rescue and use them in scaffolding. If I rerun NextClip on the same data (and run later on future data), how would you suggest to change parameters? Of course, I have consulted with the manual, I want to listen to what the thoughtful developer would say.

Very best wishes from Japan,

Shigehiro

— Reply to this email directly or view it on GitHubhttps://github.com/richardmleggett/nextclip/issues/9.

[MesutOezil]

Hi Richard,

Now replying to the previous thread about Category D. (and, my previous message to the newer thread)

I checked what proportion of the reads in Category D has residual junction adaptors. In fact, I found that it is as low as 5%. I In some cases, those reads have many low-quality bases, and it is likely that junction adaptors were not removed because of mismatches in junction adaptors. Lowering the criterion can lead not only to false matches but also to passing many low-qualoty reads to scaffolding.

I will try different parameters anyway, and see how it goes.

Best

Shigehiro

(2014/06/10 21:15), Richard Leggett wrote:

Hi Shigehiro,

Thanks for your message.

The only option in the current version of NextClip is to reduce the sensitivity of the matching, using the --strict_match (-x) parameter. However, I personally wouldn’t play around with this too much, as you risk too many false matches.

Are there that many reads that are affected? I would expect it to be relatively small.

I could think about adding a feature for a later version to handle these kind of cases.

Thanks, Richard

On 9 Jun 2014, at 06:04, MesutOezil notifications@github.com<mailto:notifications@github.com> wrote:

Hello. I have been using NextClip for these a few month and really like this program.

Let me ask a new question, the Category D. I ran NextClip on HiSeq reads (PE171) from our Nextera Mate Pair libs, and found at the end of some read in Category D, several bases probably consecutively matching the beginning of the junction adaptor. If they are genuine mates, I want to rescue and use them in scaffolding. If I rerun NextClip on the same data (and run later on future data), how would you suggest to change parameters? Of course, I have consulted with the manual, I want to listen to what the thoughtful developer would say.

Very best wishes from Japan,

Shigehiro

— Reply to this email directly or view it on GitHubhttps://github.com/richardmleggett/nextclip/issues/9.

— Reply to this email directly or view it on GitHub https://github.com/richardmleggett/nextclip/issues/9#issuecomment-45605686.