I am trying to make a really big phylogenetic tree for which I am using mega and muscle to make my 16S rRNA database and align it. Previously, I have added sequences greater than 1000 bp, aligned them and made a phylogenetic tree that turned out fine. After this, I started to collect more sequence data (based on sequence identity in NCBI BLAST) from a metagenome database called JGI/IMG. I have retrieved over 400+ sequences from this (ranging from 200-1000+ in length). Similarly, I have also downloaded 30 sequences from the SRA database.
The problem is that the sequences downloaded from these databases had a completely distorted alignment when aligned with my longer sequences , and had several massive gaps. I tried reverse complimenting the downloaded sequences and then some of them did start to align but most of them still don't. Any idea as to why these sequences don't align with my database (despite being genus or species-level matches)? How can I solve this issue?
With the description of problem it appears that you do not actually understand the basics of the process you are doing. Thus, I would highly recommend you to learn some basics of bioinformatics analysis and if possible take some related courses.
Further, alignment does not depend on where you retrieve your sequences. For an alignment the only thing which matter is if/whether the sequences are similar in identity (i.e. gene/protein), length and format.
- Format should be fasta
- You are aligning shorter sequences with longer sequences, so the problem is obvious why you are getting gaps. Plus you might be using global alignment algorithm or sequences do not have any similarity and alignment is just random
You can fix the problem by digging deeper in the problem itself.
I agree, it looks like your dataset is way to heterogeneus....you have to know that a bigger dataset does not mean a better dataset..and the proof is that when you add more but too diferent in size sequences the software begins to fail.
Furthermore, when working with mined sequences you also have to check their correct taxonomic identification (not assume so carelessly "despite being genus or species-level matches"), maybe you have some unpleasant surprises
Abhijeet Singh David Barros-Garcia I know It Doesn’t depend on the source of the data, I tried to keep the main question basic. Also, all my data is in FASTA format.
I also know that problems Can occur when aligning short and long sequences together. However, among The sequences that I downloaded, there are some that align well with the longer sequences, that is, they cover certain variable regions of the 16S completely and without any gaps, and that is how it should be, because they are short sequences. So I do understand what is going on.
The sequences that still don’t align (don’t cover a certain variable region completely) are the problem. Unless they have been published on the website itself in a scrambled way, I do not see a reason for the gaps to occur.
I understand you two are only trying to help, but I find quite rude that you assume that I “don’t understand the basics“ or that I am “careless”.
Abhijeet Singh I am trying to align short and long sequences together for my project and I figured out in the process through some testing that sequences as short as 200 bp in length, along with longer sequences, can give a reliable phylogenetic inference. hence, the conviction in adding these shorter sequences.
once a base of longer sequences has been set up, I build on it with sequences no shorter than 200. These short sequences cover different lengths of the 16S (could be any of the variable regions) continuously. Ideally, these should align with the longer 16S without any gaps. And some of them even are, perfectly in fact. I even got some good phylogenetic trees out of those. But there are some that are still completely broken up into really small chunks and they are even messing up the alignment of my longer sequences. So I don’t understand why some of the short sequences act like this while some other align without any gaps. My project supervisor is also not able to figure this out.
Now, if you still Don’t understand my exact problem, no issue. But this is for sure no basic problem.
Arguing or defending your question would not help you in solving problem. If you think I and David Barros-Garcia are trying to be rude or being disrespectful, this attitude seems immature and maybe not according to the TU Delft philosophy.
Further, if you don't consider our suggestion would be any help, I could certainly stop posting from now on.
By the way, there is for sure lack of understanding regarding phylogenetic analysis and bioinformatics. And appears that you largely misunderstood the post from David Barros-Garcia .
Samragi Chatim assuming that all the sequences you have downloaded from the databases are correctly identified at the genus/species level is risky at best. And I say this from my own experience.
Have you tried running a Genbank Blast with any of those sequences to see what results it shows?
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology