Apparently R2 and R6 are in series, and that equivalent resistance is in parallel with R4.
R2 and R6 are not in series.
However, you might want to solve this circuit using the superposition method. Then you would zero out all sources except one, solve the circuit, then choose another source, zero all the other sources, solve again, etc.
One of these steps will be to zero \$I_1\$ and \$V_1\$, and solving for the circuit response to \$V_2\$. Stop now and re-draw the circuit with these sources zero'd.
You can see that when you zero \$I_1\$, then it's effect is like an open circuit. In this step, you can now treat \$R_2\$ and \$R_6\$ as being in parallel. And when you zero \$V_1\$, its effect is like a short circuit, so now your \$(R_1+R_6)\$ is effectively in parallel with \$R_4\$.
So probably when your text suggested these combinations, they were talking just about this particular step of a solution by superposition.
In the steps where you consider the effect of \$V_1\$ and zero other sources, or the step where you consider the effect of \$I_1\$ and zero the other sources, you won't be able to make these substitutions.
It's also possible your book was just working out the problem of finding the Thevenin or Norton resistance of the circuit. To find that, you can zero out all the sources, and again the suggested combinations apply. But that is only for one step of finding the Thevenin or Norton equivalent. When you're working on finding the Thevenin or Norton source values, you won't be able to make those substitutions (except in certain steps if you solve using superposition, as discussed above).