What matters is the resultant of the force vectors for gravity and cornering force. The resultant vector has to intersect the line between the contact patches. If you were cornering at 1g, the bike would be leaned over at 45°, and the two vectors (cornering force horizontal, gravity vertical) would be equal in magnitude, and their combination would be at 45°. The same sort of thing would be true at lesser cornering forces, with the angle being different, of course.