Identification of potential crucial genes in atrial fibrillation: a bioinformatic analysis