The problem here is the misleading nature of the term “coding strand”.
I have previously posted on this topic making the point that mRNA — which is what the genetic code relates to — should be the basis of nomenclature, and that the best and least ambiguous terminology is sense (the way the code is read from the mRNA) and antisense (its reverse complement). To quote myself:
The idea seems to me clear that when you read the string of codons
that encode the amino acid sequence from this strand they ‘make
sense’. (Anti-sense is used in preference to non-sense, as ‘nonsense’
was the term used historically for mutations that converted amino acid
codons into stop codons.) It can also be extended to non-protein
coding genes (e.g. for tRNA), where ‘sense’ correlates with the
sequence of the gene product.
Thus, if you consider the two strands of DNA, the one that is transcribed by RNA polymerase into DNA is the anti-sense strand, and the other (which is the one humans may read after a DNA sequencing experiment) is the sense strand.
My main objection to the use of coding and non-coding for DNA strands (apart from logic — how does one code and the other not?) is ambiguity:
As the anti-sense strand is the template for RNA polymerase, one might
make the mental association between this and ‘coding’, whereas it is
the sense strand that is meant [by coding] in this (admittedly common) usage.
So in my opinion the question is to blame, although this is very common. But, to answer the question, one presumably should assume that the strand with the asterisk is the sense strand, and any changes in it are reflected directly (apart from T to U) in the mRNA.
Footnote
It is ironic that the genetic code provided in this question is unlikely to be correct, although the organism is not specified. For example, in vertebrates the following applies:
AGA Stop (not Arg)
AGG Stop (not Arg)
AUA Met (not Ile)
UGA Trp (not Stop)
[https://www.ncbi.nlm.nih.gov/Taxonomy/Utils/wprintgc.cgi]