Manhattan plot interpretation

In the modern system, the same physical position should not have different rs numbers for different mutations. Instead, all possible variations at that specific coordinate are grouped under a single rs ID.

Here is how the system handles the "3 possible mutations" at every spot:

1. The "One ID, Multiple Alleles" Rule

Even though a single position in the DNA (e.g., Cytosine) can mutate into three other things (Adenine, Guanine, or Thymine), the dbSNP database assigns one rs number to that specific "locus" (location).

In the database entry for that rs ID, you will see a list of Alleles. For example:

Position: Chromosome 1, 123456

ID: rs100

Alleles: C/A/G/T (This means the "Reference" is C, but they have found people with A, G, and T at this exact same spot).

2. Multi-Allelic SNPs

When more than two variations exist at one spot, scientists call it a Multi-allelic SNP.

Most SNPs are bi-allelic (only two versions exist in the population, like A or G).

Some are tri-allelic or quad-allelic.

Even if it is quad-allelic (all 4 bases are found there), it still keeps the same rs number to avoid confusion.

3. Why it was confusing in the past (Legacy Data)

In the early days of genetics, different labs discovered different mutations at the same spot and gave them different names. This created "duplicate" rs numbers for the same location.

However, the NCBI (which manages these IDs) performs "merges." If they find two rs IDs pointing to the same coordinate, they retire one and merge it into the other. This is why you will sometimes see a note saying "rs123 has been merged into rs456."

4. How this looks on your Manhattan Plot

When you see a peak on your graph labeled rs12345, that peak represents the association of that location with the trait.

It doesn't tell you which mutation (A, G, or T) is the "risk" version.

To find that out, you have to look at the Summary Statistics table, which will list the Effect Allele (the specific letter that causes the change).

https://www.youtube.com/watch?v=PFdRuCYJ4TA