Changed prediction csv column headers, added an additional key 'visualization_NGD_inverted_scores' where k=2-4 scores are inverted as 1-score for more intuitive visualization, implement check for sequence ID uniqueness