The very fact that you have found Catherine Brannigan, born in Ireland 1835 certainly is of interest. Note Patrick and Catherine’s year of birth found on the Family Search-1835. They may have been twins.
Could Peter’s or Mary’s parents have raised the children for a while?
Years of birth should be regarded as approximate (especially for Irish people of that era). They were calculated from age stated on a document. That age may not have been accurate.
Patrick was 21 on 1855 census and 24 on 1860. Both ages can't be correct. If his age was correct on 1 census, it narrows his year of birth to 2 years, depending on whether the anniversary of his birth was before or after census day. Patrick's ages on the 2 censuses gives approximate years of birth 1833-1836. His age may have been incorrect on both censuses, in which case his birth window expands further in both directions.
Wasn't Catherine's birth year calculated from information on a death registration? Even more scope for error than on a census as the subject wasn't alive to provide information.
A household return for a UK census usually listed only people who spent census night in that household. It wasn't supposed to include family members who were away from home.
1841 Scotland census was taken in June.
Information about 1841 Scotland Census:
https://www.ancestry.co.uk/search/collection/1004/Added. Link gives message "No longer available". It can be found by internet search for Scotland census 1841.