The Sign A 'water' (Labat #579) can receive infixed signs. When it does so, it splits into two parts without glyphic connection, one vertical wedge to the left of the infixed sign, and a vertical stack of two vertical wedges as its second part (Labat #585a NIMIN3, like HAL turned vertical). There is no structural justification for encoding these as a sequence of three signs, DISH, Infixed Sign, and Labat #585a. Examples can be seen in Labat ##580-584. The signs include
A x BAD
A x KAR2
A x MUSH
A x IGI
A x DU6
A x A
A x HA
A x SAG
A x HA

The signs UM (Labat #134) and DUB (Labat #138) can receive infixed signs. When they do so, they split into two parts without glyphic connection, a left part, followed by the infixed sign, then a single vertical wedge DISH. There is no structural justification for encoding these as a sequence of three signs in that way. Rather they are UM x (Infixed sign) or DUB x (Infixed sign). An example can be seen in Labat #140. The signs include
UM x U
UM x (ME.DA)
UM x (U.LAGAB)
DUB x SHA3
DUB x LAGAB
DUB x ESH2
DUB x LAGAB
DUB x SHE
As an additonal note, DUB is one vertical wedge at the end more than UM. But there is no reason to encode DUB x LAGAB as UM.LAGAB.DISH (similarly for others) Similarly, UM+ME is a traditional analysis, presumably there is some reason this has not been given as DUB.ASH?

In addition to the productivity considerations given above, best typography for Container x Infix sign types also treat the infixed signs differently by rendering them in smaller size, signaling their infixed status, EVEN when the container sign is split in the ways illustrated here, has nothing above or below them so there would be no need to reduce the size of the infixed sign component(s). Thus an attempt to encode the fragments would get the wrong glyphic result, as well as being structurally inappropriate.

There are also some sequences into which signs can be infixed. Whether this demonstrates that the "sequences" are to be treated as units for encoding I do not wish to speculate. I'm not yet convinced that such a conclusion would follow. I think we rather have here something interesting about where infixation can work.
MUNU4 = MUNU6 x SHE or also (BULUG3 x SHE)