Should I ignore these DNA matches?

[coombs]

You can get NME's as well, if a married man had a fling, and the single woman gave birth and the mans wife took the baby on as theirs. Or if the wife was "barren" and another woman offered to be the incubator for the husband and his wife, and the hubby had to do the deed with the woman to get the ball rolling.

[Gadget]

And, of course, sperm and egg donation.

[Biggles50]

Guy wrote:-

I would also suggest creating quick and dirty trees, i.e. using sources like online trees to build an unchecked data set to provide names to matches. (it goes without saying that any such use requires extra research later to prove the assumptions being made in the quick and dirty tree but it can be helpful if care is used).
As you keep triangulating and discounting maternal matches you will also build a dataset of possible paternal matches and by elimination will eventually produce pointers to the paternal lineage.
.
.
We must work in a very similar manner for this is what I also do.

Rather than clutter up my main tree with Floating Branches or unattached DNA matches and their family, I start a new tree.

ie
Surname of Match 01, I put them in and their parents and grandparents if known and rather than expand it I copy it and rename it as Surname of Match 02.

I then work on expanding the tree but if it turns out to be a brick wall I leave it, go back to the basic 01 tree, copy it again but this time name it as 03.

Then I continue as usual but hoping I am on the right track, repeating the starting over copy and rename as necessary. 

If I get it right I Tree Share the DNA matches tree with Roots Magic, then import the matches tree into my Main tree in Roots Magic and then Tree Share back to my Main tree in Ancestry.

BTW
I avoid using Thrulines, but do use Common Ancestors provided I work through each person in the suggestion line by line verifying each with the usual supporting references.