Hello, it is possible to analyze the length of haplotype (eg the combination of the SNP) to detect signal of positive selection in the genome. The general idea is that a locus under selection pressure will quickly display a long haplotype. For example see the paper entitled "Positive Natural Selection in the Human Lineage", PMID 16778047

There are multiple possibilities.

First, you can convert VCF to PED format by PLINK. This format and its binary form BED (not to confuse with BED genomics intervals format) are used quite commonly in software for SNP analysis. TASSEL (https://www.maizegenetics.net/tassel) seems to what you need, as it contains evolutionary tools.

Alternatively, if you are familiar with R, you can use packages "vcfR" and "ape". "ape" is a well known package for evolutionary biology with many tools, and "vcfR" is specific for reading and handling VCF files in R. If you have reference sequence vcfR can create new sequence with variants imputed, and it can be used further by ape or exported for external applications.

Thanks again to all!.

Dear Alexandr; what if I don't use the reference genome? I actually have extracted my gene from the whole genome. It's a gene with quite few duplication copy each differ in a few a.a but different function.

Based on the large number of SNPs , I want to see whether any of alleles of those genes/ each gene have been positively selected and sort of being fixed in the population?

So, this way I think I can exclude the reference genome and based on the SNPs I have run and evolutionary test using probably a gene model.

Also, Are there two types of BED format?

Many Thanks,

Mah

Dear Mah,

if your scope is positive selection then you need any software computing linkage disequilibrium (LD). Both PLINK and TASSEL I mentioned above are capable to this (and TASSEL actually supports direct import of VCF, so you don't have convert it to PED, my mistake). The difference is that TASSEL provides graphical interface and plotting tools, while PLINK works purely in command line and requires further use of external tool for visualization (R, for example). Figure 1 here (https://www.cell.com/cell-metabolism/pdf/S1550-4131(16)30430-2.pdf) gives an idea of how LD is related with positive selection.

You could also try Haploview, it should do the work, but I never used it because of problems with non-human data.

Regarding BED, yes, there are two totally independent formats. BED by PLINK contatins SNP data, same as PED but in binary form. Its name is formed in analogy with other text-binary format counterparts, like SAM-BAM, VCF-BCF, and so PED-BED. PLINK's BED is also always supplemented with BIM and FAM text files. Traditional BED (genomic intervals) is always text and has no supplements.

Dear Alexandr, Thanks alot for such a complete answer. I definitely will try TASSEL. So far I have tried SNPphylo with which there is some issue in its setup. My plan was to construct a phylogenetic tree of my SNPs and this way I could also have the multiple alignement file to fo the analysis with HyPhy or PAML but could not get that yet.

TASSEL seems interesting!

Many thanks,

Mah

Dear Alexandr,

I am using TASSEL for my SNP analysis; when saving the sequences, different formats can be chosen. As I prefer the Phylip format for my further analysis I was wondering if this alignment with TASSEL is reliable? Looking at the file few sites are not exactly the same but been together made me concern if this is a real alignment or just a file format. The user manual was not very helpful in this case even though it seems the phylip file is the alignment.

Many Thanks in Advance

Dear Mah,

TASSEL data representation is some sort of pseudo-alignment, because the data are SNPs, not real sequences. When you export into PHYLIP the results are concatenated SNPs. All markers have defined physical position, so all alleles in each column correspond to a particular polymorphism.