Clustering in high complexity and low complexity barcode libraries

[ga...@nyu.edu]

Hi!

I was wondering how to tune bartender for high vs low complexity libraries. I have a few libraries with 500 000 - 1 million reads that should have no more than 96 unique barcodes, but am returned 150 - 600 clusters. I've tried tuning some parameters, but only get significant reductions in number of clusters when I make them very extreme (such as -d 12). 

Additionally, I have a high complexity library with 1.8 million reads, for which I am getting about 130 000 clusters (when using parameters). However, I only expect about 75 000 unique barcodes in this library. I know I should probably adjust parameters for this library in a different way than for the low complexity libraries, but I'm not sure what the best way to go about that is, especially since I don't know the true number of barcodes in that library. 

Thank you for your help!

Cheers,

Grace Avecilla

[赵路]

Hi Grace,

I don't know how you data looks like. But the following are general rules that helps reducing spurious barcode.

For low complexity barcode library, I recommend to set option -z be -1 so that make the cluster merging decision solely based on the distance. 

For the second relative large barcode library, you can increase -z value and distance(depends on the barcode length) to reduce spurious barcodes. 

you can find these information https://github.com/LaoZZZZZ/bartender-1.1 

Based on our observation, cluster size under 3 is very hard to tell if they are true barcodes. You can set a threshold to remove low-frequent barcodes after clustering. The threshold should be chosen based on your experiment design. 

LMK if you have other questions. 

Any advice is welcomed to improve Bartender software.

Best,

Lu

--
You received this message because you are subscribed to the Google Groups "Bartender" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bartenderRandomBarcode+unsub...@googlegroups.com.
To post to this group, send email to bartenderRandomBarcode@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bartenderRandomBarcode/3b824fe8-da07-4011-ad67-ebe01d3ea9af%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Sincerely,

 

Lu