The burglary-related murder statistics are just from the FBI website. For example, here is 2015:
https://ucr.fbi.gov/crime-in-the-u.s...nship_2015.xls
This shows it is 102 in 2015. The 2017 table shows 90.
The FBI does distinguish between robbery and burglary, with the former being a violent crime aimed at a person, and burglary aimed at a structure for purposes of theft. About 16.5% of robberies occur in a residence, and, if the murder rates for robberies hold true, between burglaries and robberies there's about 200 homicides a year (though, again, the burglary numbers aren't necessarily all in residences, they include commercial burglaries).
Looking at your study, it shows about 266k victims of violent crime. Of the overall numbers (3. whatever million), that's 7% of burglaries that involve violence. Of those, MAX 200 (if you include robberies which take place in a residence and assume all burglaries are residential), end in homicides (and we don't know how many are gun homicides, remember), for a paltry 0.07% of all burglaries ending in homicides.
The math just doesn't support the idea that the threat of violent home invasions is a good reason to own a gun.
- - - Updated - - -
Edit: Nevermind that the same study you linked showed that 65% of burglaries are perpetrated by people KNOWN to the victims, which should dispel the idea of strangers invading your house is a serious problem.