Low standard error and estimated richness

[jdale03]

Hi there,

Hope I'm posting this in the right place, I haven't used GitHub very much!

I'm interested in using the Breakaway package to estimate alpha diversity in my samples, but I'm a little confused about the results I'm getting. I'm using breakaway_nof1 because my dataset doesn't contain singletons (I used DADA2 without pooling samples).

The two main issues I'm noticing:

1. the standard error is extremely small for most of my samples (basically zero)
2. the estimated richness is actually smaller than the observed richness

An example of the output I'm getting:

> ft %>% head(10)
   Var1 Freq
1     2    4
2     3   10
3     4    9
4     5    6
5     6   11
6     7    3
7     8    5
8     9   10
9    10    9
10   11    4

> ft %>% tail(10)
    Var1 Freq
99  1030    1
100 1047    1
101 1354    1
102 1575    1
103 1728    1
104 2167    1
105 2532    1
106 2612    1
107 6870    1
108 9753    1

> ft %>% sample_richness
Estimate of richness from method Plug-in:
  Estimate is 217
 with standard error 0
  Confidence interval: (217, 217)

> estimated_richness_no_singletons <- breakaway_nof1(ft)
> estimated_richness_no_singletons
Estimate of richness from method PoissonModel:
  Estimate is 116
 with standard error 0.06
  Confidence interval: (116, 116)

Most of my samples have had results like those above, but I do have a couple of samples with a much higher estimate and standard error, for example:

>ft %>% head(10)
   Var1 Freq
1     2    2
2     3    6
3     4    4
4     5    6
5     6    5
6     7    8
7     8    9
8     9    6
9    10    6
10   11    3

> ft %>% tail(10)
     Var1 Freq
136  2511    1
137  3126    1
138  3155    1
139  3161    1
140  3394    1
141  3552    1
142  4950    1
143  6339    1
144  8537    1
145 31073    1

> #observed species richness:
> ft %>% sample_richness
Estimate of richness from method Plug-in:
  Estimate is 267
 with standard error 0
  Confidence interval: (267, 267)

> #estimated richness (w/o singletons)
> estimated_richness_no_singletons <- breakaway_nof1(ft)
> estimated_richness_no_singletons
Estimate of richness from method breakaway_nof1:
  Estimate is 4439
 with standard error 293585.27
  Confidence interval: (268, 13074178)

I was wondering if you had any ideas about what's going on here? My only guess so far is that it's because the number of doubletons and other low-abundance ASVs in my data is actually pretty small (compared to the example data in the breakaway vignette where there are a large number of species observed once, twice, three times, etc). I'd appreciate any insight you can offer (including if I'm missing something obvious!). If there's any more information I can provide about my data please let me know.

Thanks for reading!
Jenna