ambiguous bases with v1-v2 region after make.contigs

We sequenced the v1-v2 region with the MiSeq platform. After putting the 2 reads together with make.contigs there were a lot of ambiguous bases created, which downstream removes many more sequences than we would like.

Is there a work around for this using mother?
Is there potential that our reads are not creating contigs very successfully because the v1-v2 region is longer than 250 bp? If so, what can be done with the current sequences using mother?


Sorry you’re struggling with this. While we didn’t discuss V12 in Kozich et al. (, we did show that your reads must fully overlap to get good noise reduction in the data. I’m afraid that unless you co-sequenced a mock community and can re-do the analysis we did, you’re stuck. I know it sounds glib, but my solution would be to not sequence the V12 region. Alternatively, you could reduce the delta q threshold in make.contigs to reduce the number of Ns, but being aware that you’re also going to significantly increase your error rate.


Thank you for your input.

We currently have other samples that have also been amplified with the V12 region that we would like to send off for sequencing. Instead of completely starting over with new primers, do you think that sequencing the V12 region using the MiSeq chemistry 2x300 would help to get rid of the overlap issues?

Thanks again for your advice

How long is your V12 amplicon? Do you have a mock community that you can sequence?

Our amplicons are 312 bp long before trimming anything. No we do not have a mock community…

Sounds like it could work then. It would be a good idea to make a mock in the future for these types of issues.

Great, thank you for your input! We are currently debating finsihing these samples with the V1-2 region and running them with MiSeq 2x300 vs buying new primers all together in the V4 region and starting over with the amplifications. If we do run the V1-2 samples with the 2x300 I will post back with updates.