I would like to create a sequence logo from my variant frequency (short 15bp DNA sequence). So it's not the normal alignment of several sequences but only ONE single sequence where I would like to show the heteroIogy I see in my sequencing reads. I have the data as an excel file at the moment but can't use this to upload at WebLogo3. I tried to save it as tab-delimited .dat file (transfac file as I understood) but when I upload it, I receive an Error that the file can't be parsed.
Has anyone experienced and overcome this?
Are there other ways to create a suitable file format for uploading at WebLogo3?
Or are there other web based programs that create sequence logos (from a frequency table)?
There are probably some kind of scripts for extraction of fasta from the xls files, but you can do it manually. Just put all your sequences in 1 excel column, the names - in another column to the left. Than you need to make a blank column to the left of the "names" column and write there ">" sign in all cells and the same with any unique sign (as, eg. #) to the right of the names, thus the data should look like:
column1 column2 column3 column4
> seq1 # agcg
> seq2 # gggg
Save it as a tab delimited file, open it in MS Word and replace tabulations with spaces while "#" sign with "add a paragraph" option. You will get a fasta file as a result.
Hej Ilya thanks for your response! I think I wasn't clear enough. I don't want to use the seq logo in the conventional way to compare several sequences but just want to show my variant frequency of eacht nucleotide at the respective position from one single sequence. So have way to show the frequency of SNPs in a heterologous sequencing product.
The sequence logo tools I am familiar with, such as WebLOGO, take a multiple sequence alignment in FASTA format (or simiar formats) and make the logo for you.
http://weblogo.threeplusone.com/manual.html
http://weblogo.threeplusone.com/create.cgi
The HIV Databases QuickAlign tool:
https://www.hiv.lanl.gov/content/sequence/QUICK_ALIGNv2/QuickAlign.html?sample_input=1
will give you the counts of each base or amino acid at each location in your alignment, as well as producing a LOGO. For organisms other than HIV, you select the "other" organism and upload your own alignment.
I am attaching a sample input alignment and outputs, so you can see if this is what you need.
Hi Brian Thomas Foley !
Thanks for your detailed response! I used weblogo so far and it works great for sequence alignments, but I would like to show the heterology of my reads just from one single sequence (I edited the original question, hopefully it gets more clear now).
If your reads are heterologous, they are not from "a single sequence" but from a population of sequences. We deal with this in the HIV world every day, because any given patient sample has what we call a "swarm" or "quasispecies" of virus variants in the thousand to millions of virus particles (or HIV-infected cells) in the sample. In the world of diploid organisms like most animals, we expect two alleles for most genes, so we should either see each base close to 100% or close to 50% one base from one allele and 50% another base from the other allele.
The sequence logos are created by first calculating the percentage of each base at each site like this
site : A : C : G : T
1 : 50 : 0 : 50 : 0
2: 100: 0 : 0: 0
3: 0: 100 : 0 : 0
etc.
and then the program draws the logo with an A and G each half size followed by A and C each full size.
So you can skip the multiple sequence alignment and counting the percentage of each base in each column of the alignment, and just use a table of percentages or counts to create a LOGO.
Brian Thomas Foley yes this is exactly what I would like to do! I also have the seq data from a virus and expect a heterologous population. I just don't know how to upload the table of percentages to a programm that then creates the logo from this.
Artur Burzynski
it is the percentage of each nucleotide at a given position (frequency of SNPs), similar to what Brian has commented. So it's not a sequence alignment and I guess fasta would not be an appropriate file. I would not mind converting my excel table into any other table/matrix, but just don't know which one or how
Helena Leinweber the "position specific scoring matrix" (PSSM file format) is accepted at the site I link below. You may need to get some help from them to get your file format right.
https://services.healthtech.dtu.dk/service.php?Seq2Logo-2.0
- Badges
- Science method