Reading right-to-left, a chemist plans a complex target molecule (right, large violet node) by retrosynthetic analysis: mentally cutting a strategic bond (a "disconnection") to reveal two simpler synthons, then repeating on each precursor until reaching cheap, commercially available starting materials (left, small blue nodes). Woodward's and Corey's classic total syntheses — reserpine, strychnine, prostaglandins — are exactly this: a tree of disconnections planned backward, then executed forward in the lab bond by bond.
overall yield = ∏ (step yield_i) for i = 1..N steps
protecting group: +2 steps, but rescues one step's yield
from side-reaction loss (shown as red byproducts)
- Synthesis steps — how many bond-forming reactions separate the starting material from the target; the retrosynthetic tree gets one node per step.
- Per-step yield — the average fraction of material carried through each individual reaction; because yields multiply, even a "good" 85% step, repeated eight times, only carries ≈27% of material to the finish.
- Reaction pace — how fast reagent particles flow along the route in forward-synthesis mode.
- Protecting groups — toggling this masks a reactive functional group (translucent cyan shell) around the synthesis's most failure-prone step, buying it a higher yield at the cost of two extra protect/deprotect steps — the central trade-off of real multi-step synthesis planning.
Switch between Retrosynthesis (dashed arrows pointing backward from target to precursors, disconnection analysis) and Forward synthesis (solid arrows, reagents actually flowing left to right, building the molecule up) to see the same route from both directions chemists use it.