Former chief scientific adviser said it was a ‘mistake’ to show public worst-case 4,000 deaths per day projection to justify second lockdown