GWAS data from PGC

this is my first time using the GWAS data, I downloaded some data from PGS and I Have some questions.

1- there are a lot of SNP to the same gene with different P-values, why does this occur? and if I want to use one value P-value to each gene? it is better to take may the minimum p-value

2- the p-values in the databases are different than the p-values mentioned in the corresponding paper, why do these or those results after QC by using Plink or MAGMA ?

3- is it possible for example to get the final results ( genes, snpid) for all bipolar disorder databases?

4- why are some genes have different P-values in different databases for the same disease?