OK.
I think we are talking about a single steel beam shaped like an upside down V? There is a beam at ridge (top of the upside down V). I assume that timber rafters span beween the ridge beam and the top of a wall and that the same wall provides the support for the steel beam under discussion! Furthermore I assume that the structure is effectively symetrical about the ridge line. Normally there would be ceiling level ties connecting the rafters - I assume that this is not the case and that the roof underside is exposed.
All the following is based on the above assumptions being true.
The steel beam will deflect under the load and cause the supporting walls to move apart. The lengths either side of the ridge will experience both bending and compression forces.
In terms of doing a design, I guess you don't have frame analysis software and will be doing a hand calculation. Initially work out the reaction forces and directions at the supports. Resolve these forces into axial and normal directions relative to the section axis. Now treat the member as a cantilever with the support reactions loading the tip and the point of fixity being the ridge. You can now determine the bending moment (maximum at ridge) axial compression in the steel section and deflection of the tip normal to the section. The deflecton can now be resolved back into horizontal movement of the wall and vertical sag of the ridge. The steel section can be designed for combined axial and bending.
Simple eh? - It's all down to geometry.