Adonis(vegan) in R with Strata

[Daiv9]

Hi,

I have paired samples, so I cannot use compare_categories.py - Adonis code given by qiime directly. I have to use R Adonis with strata function to identify paired samples as a block variable. 

Can anyone guide me how to write the code in R for that ? Can i use "unweighted_unifrac_dm.txt" directly in R Adonis package ?

I have attached both the files, distance matrix file and Map file. Strata, block variable will be Patients and I want to run Adonis to find difference between Stage (S,L)

Thanks

Daiv

[Aditya Bandla]

Hi Daiv

I propose the following code to answer your question

dist = read.table("unweighted_unifrac_dm.txt", header=T)

result = adonis(dist ~ Patients + Stage, data=S1L2)

I generally use the OTU table as the input and make use of the dist option. However, I understand vegan does not support unifrac as a distance measure. So I am a bit doubtful whether one can use the distance file directly as the dependent variable.

Further, I have experimented with the strata function in Adonis. The order of the variables in the function call is important. With the above function call, Patients will be treated as a blocking variable and thus there is no need to use the strata option. Even if you did use the strata option, the outcome would be the same

Cheers,

Aditya

[Daiv9]

Hi Aditya,

That's awesome, it worked. You made my day !!!

Though, with strata argument, it did not give me the same results. see below,

Without Strata,

Call:

adonis(formula = dist ~ Patients + Stage, data = read.table("C:\\Share\\Qiime\\S1L2Map.txt",      header = T)) 

Permutation: free

Number of permutations: 999

Terms added sequentially (first to last)

          Df SumsOfSqs MeanSqs F.Model      R2 Pr(>F)   

Patients   1    0.1943 0.19433 0.94012 0.02385  0.634   

Stage      1    0.3054 0.30537 1.47730 0.03748  0.010 **

Residuals 37    7.6483 0.20671         0.93867          

Total     39    8.1480                 1.00000          

---

Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

With Strata

Call:

adonis(formula = dist ~ Stage, data = data, strata = data$Patients) 

Blocks:  strata 

Permutation: free

Number of permutations: 999

Terms added sequentially (first to last)

          Df SumsOfSqs MeanSqs F.Model      R2 Pr(>F)  

Stage      1    0.3054 0.30537  1.4796 0.03748  0.027 *

Residuals 38    7.8426 0.20638         0.96252         

Total     39    8.1480                 1.00000         

---

Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Thanks a ton !

Daiv

[Daiv9]

Hi Aditya,

how do you use the OTU table as the input and make use of the dist option ? Do you use the otu_table.biom file or you use the summarized taxa otu table?

Best

Daiv