I am doing tumour-normal Ampliseq comprehensive cancer panel analysis on Ionreporter. I am looking for deleterious somatic mutations. For this purpose, I have designed following filter in Ionreporter-

(Variant Type in INDEL, LONGDEL, SNV, MNV
Variant Effect in stoploss, nonsense, missense, frameshiftInsertion, frameshiftDeletion, frameshiftBlockSubstitution
0.0 <= PValue <= 0.05
10 <= Allele Read-Count <= 100000
0.0 <= SIFT <= 0.05 OR 0.95 <= PolyPhen <= 1.0)

Using this filter, I am getting 10 mutations in one tumor-normal pair, in other 3 pairs I am getting fewer mutations (2-3 only in each pair) because of lesser coverage and coverage uniformity in pairs. Mutant allele frequency (allele ratio) ranges from 0.03 to 0.44

Please suggest me how I can find mutations with confidence? And how may I have an idea about sample purity?

Thanks in advance