"I just changed methyl to ethyl..."
Sena was confused. Activity dropped to one-tenth.
"It was too heavy," Akira said matter-of-factly.
"But it's just one more carbon atom?"
Lina opened her laptop. "Let me show you."
The screen displayed the protein binding pocket. The original compound fit nicely.
"With methyl group," Lina explained. "Fits perfectly in the pocket."
Next, the model with ethyl group appeared. The substituent clashed with the pocket wall.
"Oh..."
"Binding pockets have limited space," Akira said. "Ethyl group causes steric clash."
"This clash reduces binding energy," Lina supplemented.
Sena rotated the model. "So propyl group would be even worse?"
"Obviously," Akira answered immediately. "Completely protrudes."
"But," Sena thought. "What about methoxy? Similar size but different."
Akira narrowed his eyes. "Interesting question."
Lina quickly created the model. Methoxy group. Contains oxygen.
"It fits," Sena was surprised.
"Not just size. Electronic properties matter," Akira explained.
"Methyl is electron-donating. Methoxy is also electron-donating but polar."
"Polar?"
"Oxygen is highly electronegative. So it can interact with hydrogen bond acceptors in the pocket."
Lina displayed the interaction. The methoxy oxygen formed a hydrogen bond with a serine residue.
"This increases activity."
Sena compared structures. "Substituent choice is really important."
"The core of SAR, structure-activity relationship," Akira said. "Small changes have big impacts."
"What about fluorine?" Sena asked. "It's small, right?"
"Fluorine is special," Akira became serious. "Highest electronegativity. Strongly electron-withdrawing."
"Favorable for binding?"
"Depends. Fluorine often increases metabolic stability."
Lina supplemented. "C-F bond is strong. Hard for metabolic enzymes to cleave."
"But effect on activity?"
Akira answered carefully. "Position-dependent. Some positions increase activity, others decrease it. So we test each."
"Test everything?"
"Try various substituents at all possible positions. That's medicinal chemistry."
Sena was overwhelmed. "Infinite combinations..."
"That's why strategy is needed," Akira emphasized. "Don't change randomly. Form hypotheses."
Lina displayed a table on screen. "Methyl, ethyl, isopropyl. Fluorine, chlorine, bromine. Methoxy, ethoxy."
"Consider both steric and electronic effects, prioritize."
"First check if size fits. Then evaluate if electronic properties are favorable."
Sena understood. "Logical, not random."
"Yes. But unexpected results are common," Akira admitted. "That's why experiments are needed."
"Can't just calculate?"
"Calculation shows direction. But ultimately, synthesize and measure."
Lina displayed other data. "This is my QSAR model. Trained on substituent-activity relationships."
"QSAR?"
"Quantitative Structure-Activity Relationship. Machine learning to predict activity from structure."
A graph appeared. Predicted and observed values nearly matched.
"Amazing..."
"But not perfect," Lina was cautious. "Poor at extrapolation. Completely new substituents can give wrong predictions."
Akira summarized. "Substituent selection is fusion of science and experience. Theory, computation, experiment. Use all."
"The judgment that this substituent is too heavy"
"Results from considering steric effects, electronic effects, metabolism, solubility. Everything."
Sena gazed at the structure. One small substituent. But its selection had deep meaning.
"What substituent should we try next?"
Akira thought. "Cyano group. Small and electron-withdrawing. From pocket shape, might work."
Lina started modeling. "Let's try."
Small substitutions make big differences. That was the fascination of drug design.