Decontam question

[mcavoytj]

I have run decontam on a data set that has ASV’s for tumor tissue, negative controls, and control tissue. The negative controls did not have any tissue but only reagent. The controls are non-tumor specimens from the same body location as the tumor specimens. I have run tumor-negative controls and controls negative controls. I am wondering whether if I run tumor versus controls (considered as contaminants) it will help to distinguish control ASVs from tumor ASVs?

[benjjneb]

I am wondering whether if I run tumor versus controls (considered as contaminants) it will help to distinguish control ASVs from tumor ASVs?

(Assumption, "controls" in the above sentence meant non-tumor specimens): No, this is not recommended. decontam is intended as a tool to identify and remove contamination in the sense of microbial reads that are in the data that weren't in the sample, and should be use with "negative controls", i.e. samples that should have no true sample DNA.

What you are considering is contrasting "case/control" samples to look for the relevant differences, which requires a different analytic approach.

[mcavoytj]

Hi Benjamin: Many thanks for your reply. What analytic approaches would you recommend to detect relevant differences between controls and tumor samples? We have not looked at any approaches so far. Thanks again. Tom

On Sat, Oct 17, 2020 at 4:23 PM Benjamin Callahan [email protected] wrote:

I am wondering whether if I run tumor versus controls (considered as contaminants) it will help to distinguish control ASVs from tumor ASVs?

(Assumption, "controls" in the above sentence meant non-tumor specimens): No, this is not recommended. decontam is intended as a tool to identify and remove contamination in the sense of microbial reads that are in the data that weren't in the sample, and should be use with "negative controls", i.e. samples that should have no true sample DNA.

What you are considering is contrasting "case/control" samples to look for the relevant differences, which requires a different analytic approach.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/benjjneb/decontam/issues/82#issuecomment-711075041, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARM3IEXWHUM3OHCTO2PVOQDSLH4KTANCNFSM4STPPMKA .

[benjjneb]

I am reticent to make any recommendations about what community-wide dissimilarity of differential abundance methods you should use on the DADA2 site.

Here we care greatly about helping you get an accurate and useful description of your sample from your amplicon sequencing data. The analysis fo that description is probably best conducted using some tools outside of DADA2, perhaps including Permanova, Phyloseq, and perhaps the vast universe of other methods that area available via the R statistical analysis environment.

[mcavoytj]

Thanks. I will check the tools out. Tom

On Sat, Oct 17, 2020 at 9:04 PM Benjamin Callahan [email protected] wrote:

I am reticent to make any recommendations about what community-wide dissimilarity of differential abundance methods you should use on the DADA2 site.

Here we care greatly about helping you get an accurate and useful description of your sample from your amplicon sequencing data. The analysis fo that description is probably best conducted using some tools outside of DADA2, perhaps including Permanova, Phyloseq, and perhaps the vast universe of other methods that area available via the R statistical analysis environment.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/benjjneb/decontam/issues/82#issuecomment-711101062, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARM3IEQXKEDF4UBL3KJX56TSLI5IJANCNFSM4STPPMKA .