When you have a paired design so that you can match each Ab+ to a specific Ab- control, there is no need of error bars for controls.

pnas.org/content/early/… More discussions on negetive control error bars: researchgate.net/post/qPCR_Stan…

pnas.org/content/early/… There‘s no need of error bars for CKs if you have a paired design, in which you match each Ab+ to a specific Ab- CK