genePredToGtf cannot work, as it convertes the genePred format to GTF so the input must be in genePred format. The dbSNP tables are not in genePred format as they are not gene predictions. GTF is a bad match for this type of data, REDItools made a weird choice, as GTF has fields for exons and frames that don't make sense for SNPs. Either way you can use awk to convert the file to a very bare-bones GTF file which is probably good enough for this software:
Its seems work, Thanks for help, I also think that using GTF format to complete these filtering steps is strange, so I am considering whether to write a program to use VCF or BED files to perform the filtering work on the results.
Gtf is usually for genes. It’s not a very efficient solution for SNPs. Which software requires gtf for SNPs ?
REDItools, a software for detect RNA editing, see:
https://github.com/BioinfoUNIBA/REDItools/blob/master/README_1.md#filtertablepy
for more detail