Values of S.chao1 & ACE bigger than total_counts

[lynneitelson]

Hi,

I have a table of OTU's and how many times they were seen on each media:

I used the code written below and got strange results. The statistical values of ACE and Chao1 are larger than the total counts (total counts was calculated in the initial excel work sheet by applying =sum to each column).

What might be wrong?

The results:

Thank you!

The code:

`library("vegan")

#=============upload and fix the table=================#
media_table <- read.csv("Media.csv")

media_table <- subset(media_table, select = ISP2:MA)
media_table$MA <- media_table$ma

#=================Calculations of chao ace & sobs==================#

community <- as.matrix(media_table)

transposed_matrix <- t(community)
chao <- estimateR(transposed_matrix, "chao")
df_chao <- as.data.frame(chao)

write.csv(df_chao1, file = "X:/MichaI_Shared/users/lynne/Experiment/cultivation/chao+ace/df_chao_ace_sobs.csv")
`

[jarioksa]

What looks like wrong? Your results cannot be found from your data table, and I have no idea how you got those results, and what is wrong with them. However, the Chao and ACE values can be larger than the total count – after all, they are estimates of unseen species, and what is unseen is unseen also to total count. I think that for most data sets this happens rarely, but if none of the OTUs has very high counts, and there are many OTUs with count 1 the estimates with unseen species can be much higher than the total count.

The estimated number of unseen species, or the number of species added to the observed number of species is found from the number of species seen only once (singletons) a₁ or twice a₂ in estimateR as a₁ (a₁ – 1) / (a₂ + 1) / 2. It is easy to see that this can be a large number if there are many OTUs that were seen only once, or a₁ is large and a₂ = 0 or small.