Question

Sparsity plot and how that indicates a model fit to negative binomial distribution

0

Entering edit mode

tg369 ▴ 40

@tg369-13587

Last seen 10 months ago

United Kingdom

I learn from DESeq2 that sparsity plot can provide insight regarding a data fit to negative binomial.

Running my RNA seq data (3 controls and 3 tests) in DESeq2 and then generating sparsity plot using

plotSparsity(), I got a graph (https://www.dropbox.com/sh/bj12eelleqzehzx/AAD8MHoNIRhueXTdWyyl_ckqa?dl=0). Please could

you explain me how that graph should look like and advise me if it would be OK to
go ahead with DESEq2 for this data?

I highly appreciate your explanation and very much looking forward to that. Thanks, Tanay

deseq2 • 2.0k views

ADD COMMENT • link 7.4 years ago tg369 ▴ 40

score 2 · Answer 1 · 2017-07-27

2

Entering edit mode

Michael Love 43k

@mikelove

Last seen 7 hours ago

United States

I forgot, I had the image in my email from previously. This looks fine given the sample size of 3 vs 3. It means that for some genes all of the counts are from one sample but this is concentrated for the low to middle count genes, and the NB can accommodate this with higher dispersion values.

ADD COMMENT • link 7.4 years ago Michael Love 43k

Wolfgang Huber · Answer 2 · 2017-07-27

1

Entering edit mode

Michael Love 43k

@mikelove

Last seen 7 hours ago

United States

Sorry I'm away from WiFi and can't access the PDF, I'll take a look as soon as I can. Generally I was using the plot as a diagnostic for when the majority of genes with high counts have most of the row sum coming from individual samples.

ADD COMMENT • link updated 7.4 years ago by Wolfgang Huber ★ 13k • written 7.4 years ago by Michael Love 43k

score 0 · Answer 3 · 2017-07-27

0

Entering edit mode

tg369 ▴ 40

@tg369-13587

Last seen 10 months ago

United Kingdom

thank you Michael for this helpful explanation.

ADD COMMENT • link 7.4 years ago tg369 ▴ 40