Does M31 result from an ancient major merger?

The numerous streams in the M31 halo are currently assumed to be due to multiple minor mergers. Here we use the GADGET2 simulation code to test whether M31 could have experienced a major merger in its past history. It results that a 3+/-0.5:1 gaseous rich merger with r(per)=25+/-5 kpc and a polar orbit can explain many properties of M31 and of its halo. The interaction and the fusion may have begun 8.75+/-0.35 Gyr and 5.5 +/-0.5 Gyr ago, respectively. With an almost quiescent star formation history before the fusion we retrieve fractions of bulge, thin and thick disks as well as relative fractions of intermediate age and old stars in both the thick disk and the Giant Stream. The Giant Stream is caused by returning stars from a tidal tail previously stripped from the satellite prior to the fusion. These returning stars are trapped into elliptical orbits or loops for almost a Hubble time period. Large loops are also predicted and they scale rather well with the recently discovered features in the M31 outskirts. We demonstrate that a single merger could explain first-order (intensity and size), morphological and kinematical properties of the disk, thick disk, bulge and streams in the halo of M31, as well as the distribution of stellar ages, and perhaps metallicities. It challenges scenarios assuming one minor merger per feature in the disk (10 kpc ring) or at the outskirts (numerous streams & thick disk). Further constraints will help to properly evaluate the impact of such a major event to the Local Group.