prefer reference_species (e.g., mouse) over human fallback when target species is unavailable

[zhikangzhang]

Hi, @ElliotXie
 I noticed the current fallback behavior in run_complete_analysis:

 Since rabbit is closer to mouse than to human, when target_species data is unavailable, would it more reasonable to fallback to reference_species (e.g., mouse) before defaulting to human.