dist1 behaviour without unique read

julientap

unread,

Dec 19, 2020, 9:01:44 AM12/19/20

to ngl...@googlegroups.com

Hi,

dist1 option distribute shared mapped read using unique mapped read weight.

What about if there is no unique mapping read?

My guess is that in that case NGless switch to 1overN option.

Can you confirm?

++

Ju

Luis Pedro Coelho

unread,

Dec 20, 2020, 6:50:54 AM12/20/20

to NGLess List

Yes, that is correct

HTH

Luis

Luis Pedro Coelho | Fudan University | http://luispedro.org

https://orcid.org/0000-0002-9280-7885

--
You received this message because you are subscribed to the Google Groups "NGLess" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ngless+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ngless/5fde07c7.1c69fb81.d9297.13d4SMTPIN_ADDED_MISSING%40gmr-mx.google.com.

Florian Plaza Onate

unread,

Dec 23, 2020, 6:47:19 AM12/23/20

to NGLess

Florian Plaza Onate

unread,

Dec 23, 2020, 6:47:19 AM12/23/20

to NGLess

Hi Luis,

With a different tool that implement similar counting strategy, I found out that dist1 might create many false positive.

Indeed, only 1%-2% of the reads are aligned to genes without unique counts but these reads drastically increase gene richness compare to the "unique_only" strategy (+10% to +15%).

In gene abundance tables generated with dist1, I was considering setting to zero all the genes without unique counts.

What do you think about it?

Best,

Florian

Le dimanche 20 décembre 2020 à 12:50:54 UTC+1, lu...@luispedro.org a écrit :

Luis Pedro Coelho

unread,

Dec 23, 2020, 6:56:32 AM12/23/20

to NGLess List

We have done the same in the past, so it can be a good idea if accurate detections are important to you. As you write, for abundance-based analysis, this is just a tiny bit of noise, but if you are looking at detection, then you should use {unique_only} or combine the two as you propose (we could potentially implement it as an extra strategy if others find it useful).

Best,

Luis

Luis Pedro Coelho | Fudan University | http://luispedro.org

https://orcid.org/0000-0002-9280-7885

To view this discussion on the web visit https://groups.google.com/d/msgid/ngless/f0fb45d9-3405-4482-b68f-15f4918b6f78n%40googlegroups.com.

Florian Plaza Onate

unread,

Dec 23, 2020, 8:19:42 AM12/23/20

to NGLess

Thanks Luis.

I will combine both strategies then.

Happy Christmas!

Florian

Reply all

Reply to author

Forward