Identifying statistically significant chromatin contacts from Hi-C data with FitHiC2