Ignore division by 0 warnings in computing fisher, as it still correctly produces a combined p-value of 0

[bruceamm]

Output contains combined p-values = 0 in correct cases without 'divide by zero' errors (no errors or issues, AB/IS)