How to compute Fst from SNP genomic data

[@mohammadj.shamim9342]

Hello professor. I have solved the challenge of family automation as follow: breed1 = "ABR" breed2 = "ALB" write.table(breed1, 'breed1.txt', row.names = F, col.names = F, quote =F, fileEncoding = "UTF-8") write.table(breed2, 'breed2.txt', row.names = F, col.names = F, quote =F, fileEncoding = "UTF-8") # if one wants more families in one text file then extend the string as "ABR ALB" and so on. the " " simply means go to the next line If one thinks the text is going to be too much in case they do it for 100s of families, then use the loop: families = c("ABR", 'ALB', "WGA") # list of families for (fam in families){ #loops through all families write.table(fam, str_c("breed_",fam, ".txt"), row.names = F, col.names = F, quote = F, fileEncoding = 'UTF-8') print(str_c("breed_",fam,'.txt has been created')) # lets us know of created files } Of course, It is also possible to automate the whole process so that fst is calculated between all pairs of families with one click

[@katherinotalora3543]

How to calculate genetic diversity parameters, including nucleotide diversity (π), allelic richness (Ar), observed (Ho) and expected (He) heterozygosities and inbreeding coefficient (FIS), but with rarefaction process. Thank you Do you have any video of this? Thanks.

[@GenomicsBootCamp]

Hi, currently not, but these are good topics for the future. The Ho, He, and F could be computed with the --het PLINK option.

[@catarinadebettencourteavil7703]

I need the sameee

[@khuramrazzaq2388]

Hi. I have SNP data of 250 varieties that belong to 5 groups please guide me on how to find pair-wise FST and I need to see variation percentages among groups and within groups?

[@GenomicsBootCamp]

Hi, You can extract two populations all the time and run Fst with PLINK as desribed in this video, or in a more straightforward way compute a matrix of pairwise Fsts, as described in this video on the channel: Fst matrix with confidence intervals kzbin.info/www/bejne/nGi9f4d_m5uBjZY

[@GenomicsBootCamp]

For the variation you need to get the individual Fst values, and compute with the statistical software of your preference. It is e.g. the "var()" function in R

[@georgewanjala4605]

I have tried GAL and ALP, then GAL and KAR, the manhattan plot gives a similar trend.

[@mahermaaz8110]

Great sir

[@matamatosa8898]

How to calculate it for average snps? One fst for pop

[@GenomicsBootCamp]

Hi, - if you want just a plain mean Fst between the two breeds/populations, this is given by #PLINK as shown at the time point 10:42 of the video (last two lines of PLINK output) - if you want more sophisticated statistics or visualizations on it, it is best to read it into R or another environment and do it there

[@matamatosa8898]

@@GenomicsBootCamp Hi , i'm trying to do pairwise-fst between pops , i found this function in plink2 , www.cog-genomics.org/plink/2.0/basic_stats#fst but i can't get the command to work , can you tell how the command is written correctly and what constitute the first "categorical or binary phenotype name" ?

[@mahboobezamani4767]

Could you make a video using HyPhY and FUBAR?

[@GenomicsBootCamp]

Hi! Both of these seem to be useful for detecting selection signatures. I am not familiar with them though, so the first step is to get reliable results for myself. It is unclear when would that happen (time limitations).

[@mahboobezamani4767]

@@GenomicsBootCamp sure, totally understandable.

[@fakharunnisa2178]

how to increase log value on Y Axis?

[@GenomicsBootCamp]

This question is unclear. Could you re-formulate?

[@darwin6883]

you are amazing!

[@mahboobezamani4767]

How about more than two populations? I do have SNPs for 802 populations and am looking for the positive selection. Could you please advice me. Thanks

[@mahboobezamani4767]

I mean 802 individual plants to compare

[@GenomicsBootCamp]

@@mahboobezamani4767 Hi, the Fst is essentially a comparison of allele frequencies between two groups. So if you can create certain groups with your 802 plants, e.g. according to the (sub)species, lines, region of origin, or similar, you can compare these with Fst. If you have more than two groups, you can do a pairwise comparison (there is also a video for this on the channel, using the --family option) Fst can identify the differing places on the genome, but it does not say what is the reason, or if this is a positive selection, or in which of the two groups. it merely says that there is a difference. But if you can identify the genes located in the signals, it helps you is some way.

[@mahboobezamani4767]

@@GenomicsBootCamp Hi, Thanks a lot for the quick reply. I don't think I can divide them in groups and do pairwise comparison but sure I will check your other videos. Basically I do have the SNPs (vcf format) for few genes where they are taken from 802 individual plants in one single project. I wanna see where there is an allele/s in any of those genes under positive selection. My understanding is to use different evolutionary tests to make sure the positive selection is not overestimated. I thought Tajima's and FST as well as calculating π might be helpful? Or using some software.

[@GenomicsBootCamp]

@@mahboobezamani4767 Fst compares allele frequencies, so for that you need two groups. You need to find a method for a selection signature analysis that could be used within a single population. Tajima's D seems to be ok for this, but I do not know any software on this I can recommend from own experience.

[@mahboobezamani4767]

@@GenomicsBootCamp I see. It makes sense. Thank you. Will see what I can do.

[@darwin6883]

Do you have a github page for the R code?

[@GenomicsBootCamp]

Thanks for the reminder on this... The code is now available via the link in the description. Not GitHub, but another site for sharing scripts.

[@darwin6883]

​@@GenomicsBootCamp Thank you so much. I am in an evolutionary biology PhD program (you can see the inspiration for my youtube name, haha) and your videos are helping me immensely with my work.