liftover bed not matching panel gene names

Question

[whb]

I have a targeted sequencing dataset. the company provided bed file that is based on hg19. I aligned the data to hg38 (Want to be consistent in the manuscript as all other datsets from a different targeted panel were aligned to hg38) and used liftover to turn the hg19 bed file to hg38 using https://genome.ucsc.edu/cgi-bin/hgLiftOver. Then I use bedtool intersect to filter the vcf file using the bed file output from liftover.

bedtools intersect -wa -wa -a bed.file -b vcf.file

The problem is after intersect there are some genes that are not in the targeted gene list. Below is an example. Both of these genes are not targeted by the panel.

Chr hg38Start   hg38End hg19_pos    Chr SNP Gene    coverage
 chr4   139698109   139698229    chr4:140619264-140619383    chr4   139698118   H3P16   3048
 chr4   1806596        1807747   chr4:1808324-1809474    chr4   1806974 LETM1   696

looks like chr4:1808324-1809474 in hg19 correspond to FGFR3 but liftover to hg38 became LETM1? Could you please tell me what I have done wrong please?

Thank you.

[whb]

I looked at the bed coordinates again and it seems like the liftover converted it correctly (i compared it in IGV). chr4 1806596 in hg38 is FGFR3 and chr4:1808324-1809474 in hg19 is also FGFR3. Maybe Snpeff is annotating it wrong? I used the hg38 genome provided by GATK and the command line I used for snpeff was

cat $meta| java '-'Xmx8g -jar ${snpeff} ann \ '-'stats ${outdrive}/{}.mutect.snpeff.html hg38  \ ${input}/{}.mutect.pass.vcf '>' ${outdrive}/{}.mutect.snpeff.hg38.vcf