ASV of unexpected size

Commands in mothur
1

Hi mothur forum, my shared file has too many sequences in the largest ASV. The summary file looks fine. I am using mothur version 1.48.6, a PC and Windows. I am following the MiSeq SOP ( MiSeq SOP ). I make the ASVs by make.shared and classify.otu. Any suggestions for a solution?

2

Hi there - I’m not sure what you mean by “too many sequences in the largest ASV”. How do you know it’s an error?

Pat

3

Hi, thanks for the quick answer. The ASV summary file distributes sequences downwards through many ASVs. It looks similar to the shared file I made from OTUs. However, the ASV shared file distributes sequences downwards through only three ASVs. The most abundant ASV = 25 sequences, second most abundant ASV = 2 sequences, and the may remaining ASVs has only 1 sequence each. I opened the shared file with “split” and “transpose” because it was too wide (too many ASVs) for Excel to open the usual way. Perhaps there is something I have not understood.

4

Hi

You say there are only three ASVs but seem to be listing abundances for more than 3 ASVs and there are so many that it can’t be opened in Excel. Can you attach the ASV shared file to this thread?

Thanks,
Pat

5

Hi, trying to attach the ASV shared file I got this error “Sorry, the file you are trying to upload is not authorized (authorized extensions: jpg, jpeg, png, gif).”

6

Feel free to email it to me at pschloss@umich.edu

7

Thanks for sending along the shared file. I’m not entirely sure what you are asking. When I read the data into R and look at various counts I get the following…

Here is the number of sequences per sample…

   Group        n
   <chr>    <dbl>
 1 E_1         17
 2 E_2         40
 3 E_3         22
 4 E_4         17
 5 E_5         28
 6 E_6          9
 7 E_7         42
 8 E_8         68
 9 E_9         69
10 E_SW1    10307
11 E_SW2      631
12 E_negExt    76
13 E_negPCR    89
14 F_1      14761
15 F_2      26962
16 F_N       4528
17 F_P      99488

And here is the total for the 10 most abundant ASVs…

   name         n
   <chr>    <dbl>
 1 ASV00001 18837
 2 ASV00002  8572
 3 ASV00003  6063
 4 ASV00004  5753
 5 ASV00005  4828
 6 ASV00006  3920
 7 ASV00007  3875
 8 ASV00008  3349
 9 ASV00009  2282
10 ASV00010  2263

There is a wide variation in the number of sequences per sample (the first table), but I suspect that is more likely due to library generation. The first ASV is also ~2 times larger than the second, but I feel like that is pretty typical and perhaps lower than we typically see.

I would wonder why you think there are too many sequences in the first ASV or what you’re seeing to indicate that there’s a problem. Have you processed the data using OTUs or phyloytpes and seen if the results make sense at those levels?

Pat

8

Hi,

thanks. By the first ASV I mean the ASV with the highest number of sequences from the nine samples E_1 to E_9 summarized. I find that is ASV00403, with 25 sequences. Ranked by decreasing number of sequences the nex ASV has only 2 sequences. All remaining ASV has only 1 sequence each. Looking at taxlevel 6 of the summary file most sequences is found in the same genus (Planococcus). The taxonomy file assign this taxon to ASV00403, and the number of sequences is 26. However, ranked by decreasing number of sequences the next taxon has 14 sequences, followed by taxa of these number of sequences; 11, 9, 8, 8, 7, 7,… I also did OTUs and found sequences of OTUs to have a distribution comparable to the ASV summary file. I wonder why my ASV shared file seems to have only one ASV with more than 2 sequences from these E_1 to E_9 samples summarized.

Yes, there is a wide variation in the number of sequences per sample. We saw from agarose gel electrophoresis that those E_1 to E_9 amplicons had less DNA. But, all samples were sequenced simultaneously. I think a separate run for amplicons with less DNA may have resulted in more sequences from E_1 to E_9.

Kind regards,
Sigmund

ons. 16. sep. 2026 kl. 14:59 skrev Pat Schloss <notifications@mothur.discoursemail.com>:

9

The number of sequences per ASV will be less than the number of sequences per OTU and then that less than the number per genus. That’s because ASV’s are a more fine-scale resolution that those other levels. I think you’re seeing the low distribution because those E_ samples don’t have many sequences. They have between 9 and 69 sequences each (N=312 total). So, you aren’t likely to get much more than what you’re seeing.

Pat

10

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.