MSA with 80,000 30bp seqences, provides 1500 base in output file

[thh32]

I have been trying to do a multiple sequence alignment with 80,000 sequences to identify any similarities. All bases are 36-37bp long however when I run the multiple sequence alignment (tried clustal-omega and muscle however muscle failed) the output file provides the following as an example:

>NC_000853_1_10|NC_023151_4_32
-----------TA---------TACG---T-T-G-TA-GA-AAT-C--GCA-A-A-G---
G---T---G-G-T-GA--TG-T-TA-----------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
--------------------------------------



I was wondering if anyone else has had a similar issue and what might be causing this issue?

Any advice on tools which can deal with such data and provide correct output would be appreciated.

Thanks,
Tom

[GenoMax]

Tom: Are these sequences really similar (is this a region that was captured/amplified)? 36-37 bp is not that long (and ~ 80K sequences are too many) and any multiple sequence alignment software will try to do its best to somehow cram them together into a consensus leading to the result you have.

What exactly are you trying to do? Have the sequences been de-duplicated before alignment?

[thh32]

The sequences have had any duplicates removed and they are a short region which occur in a defence mechanism in bacteria but we were wondering if they have any similarity to each other. I figured it might be difficult to run but the fact it is running but providing this increased length back just seems odd.

[GenoMax]

Most MSA programs were designed before advent of NGS and using 80K sequences may be beyond the original design spec.

Have you tried the MSA in samples of 1K reads, to see if it works better? Have you considered trying de novo assembly (you may need to sub-sample if 80K reads leads to strange assembly since you haven't said how big is the region you are sequencing)? You could then try to put the individual consensus sequences from sub-assemblies together in a MSA.