{"paper_id":"49f6219b-4ea1-4631-80d9-dbabcdb6234c","body_text":"Toward a Random Background for Ligand Optimization 1 \n 2 \n 3 \nXinyu Xu1†, Olivier Mailhot1†, Galen J. Correy2†, XP Huang3†, Joao Braz4, Da Shi5, Karthik 4 \nSrinivasan1, Kara Zielinski2, Yuliia Holota6, Yuliia Kuziv6, Christos Tsoutsouvas7, Nathan 5 \nLevinzon8, Yagmur U. Doruk9, Moira Rachman1, Morgan Diolaiti9, Maisie Stevens9,Fangyu 6 \nLiu1, Katie Holland1, Harald Hü bner10, Jing Wang3, Yujin Wu1, Alan Ashworth9, Alexander 7 \nMakriyannis7, Yuqi Zhang5, Yurii Moroz11, Peter Gmeiner10, Robert Abel5, Aashish Manglik1, 8 \nAllan I. Basbaum4, Bryan L. Roth3*, James S. Fraser2*, & Brian K. Shoichet1* 9 \n 10 \n† Contributed equally. 11 \n* Corresponding authors: bryan_roth@med.unc.edu; jaimefraser@gmail.com; bshoichet@gmail.com 12 \n1Department of Pharmaceutical Chemistry, University of California, San Francisco, San Francisco, CA 13 \n94143, USA. 14 \n2Department of Bioengineering and Therapeutic Sciences, University of California San Francisco, San 15 \nFrancisco, CA 94158, USA. 16 \n3Department of Pharmacology, NIMH Psychoactive Drug Screening Program, School of Medicine, 17 \nUniversity of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA 18 \n4Department of Anatomy, University of California, San Francisco, San Francisco, CA 94143, USA  19 \n5 Schrödinger Inc, San Diego, CA, USA; 2Schrödinger Inc, New York, NY, USA 20 \n6 Bienta; Winston Churchill St. 78, 02094 Kyiv,Ukraine. 21 \n7 Center for Drug Discovery, Department of Pharmaceutical Sciences, Northeastern University, Boston, 22 \nMA, USA. 23 \n8Department of Medicinal Chemistry, College of Pharmacy, University of Utah, 2000 East 30 South 24 \nSkaggs 306, Salt Lake City, Utah 84112, United States. 25 \n9Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, CA 26 \n94158, USA. 27 \n10 Department of Chemistry and Pharmacy, Medicinal Chemistry, Friedrich-Alexander-Universität 28 \nErlangen-Nürnberg, Erlangen, Germany. 29 \n11 Chemspace LLC, Kyiv 02094, Ukraine.Taras Shevchenko National University of Kyiv, Kyiv 01601, 30 \nUkraine. Enamine Ltd., Kyiv 02094, Ukraine. 31 \n 32 \n 33 \n 34 \n 35 \n 36 \n 37 \n 38 \n  39 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 2 \nAbstract 40 \nLigand optimization is central to drug discovery  as hundreds of analogs might be 41 \ndesigned and synthesized between an initial hit and a therapeutic candidate.  The efficiency of 42 \nthis process is unclear, at least partly because there is no random background for optimization 43 \nagainst which to compare.  Such a random bac kground might emerge from synthetically 44 \naccessible but otherwise systematic random  small substitutions across starting ligands, 45 \nmeasuring likelihood of achieving a substantial improvement in affinity/potency or other property 46 \nby any single perturbation.  Recent literature and ligand-affinity/potency databases suggest that 47 \nperhaps 10% of analogs with minor modifications improve upon a parent’s potency substantially 48 \n(by >10-fold), but this number is clouded by reporting bias, intentional improvement, and inter-49 \ngroup reproducibility.  To begin to establish a background expectation for ligand optimization, 50 \nwe comprehensively and systematically modified 18 lead molecules across six targets with 51 \nsingle atom changes; 257 compounds were synthesized .  Unexpectedly, 1 1.2% of these 52 \nrandom small perturbation analogs improved potency by >10-fold over their parents.  53 \nConversely, these more potent analogs  typically had worse  in vitro pharmacokinetics (e.g. 54 \nreduced metabolic stability, lower plasma free fraction).   While it was possible to find analogs 55 \nwhere the potency increase compensated for inferior exposure and half -life, resulting in more 56 \npotent compounds in vivo, overall a frustrated landscape for ligand optimization  is revealed. 57 \nThis study begins to establi sh a background expectation for ligand potency optimization and 58 \noffers a simple strategy to do so. It also begins to quantify the challenges confronting the field 59 \nin moving beyond in vitro potency.    60 \n  61 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 3 \nIntroduction 62 \nLigand optimization is central to chemical probe and drug discovery, as initial active 63 \nmolecules rarely have the potency or  pharmacokinetic properties to be viable in vivo 1,2.  64 \nAccordingly, between the discovery of an initial hit and a clinical candidate many hundreds, 65 \noccasionally thousands of optimized analogs might be synthesized1,3.  Several strategies4-6 7,8 66 \nhave been developed to improve this process , ranging from early empirical approaches such 67 \nas Topliss Trees9 and Hansch QSAR10 to contemporary AI-guided ADMET11 prediction and free 68 \nenergy calculations 12,13. How efficient these strategies are remains uncertain, as there is no 69 \nrandom background against which to compare them.  If we use a ten-fold improvement between 70 \nanalog and parent as a benchmark for substantial impact, then only about 10% of analogs meet 71 \nthis standard in the ChEMBL ligand -protein activity database (Extended Data Figure 1 a)14.  72 \nSince these ChEMBL results suffer from success bias, sample multiple types of perturbation, 73 \nand struggle with inter -group irreproducibility15, the public domain offers no sure guidance on 74 \nwhat level of improvement one might expect in ligand optimization if one were making random 75 \nconservative changes.   76 \nRandom backgrounds have long been used in biology to help quantify significance. In 77 \ngenomics, they help to distinguish between artificial and natural selection 16. In epidemiology, 78 \nrandom incidence rates help distinguish genetic diseases fr om those driven by environmental 79 \nfactors17,18. In protein engineering, random backgrounds are used to evaluate improvements in 80 \nenzyme function 19, and alanine scanning has been used to introduce a minimally -biased 81 \nperturbation to find hot spots for ligand binding and protein func tion20. A random background 82 \nfor ligand optimization, conceivably, could quantify progress across chemotypes and targets 83 \nand compare different optimization strategies.   84 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 4 \nAkin to alanine scanning, an unbiased background for ligand optimization might involve 85 \nconservative perturbations, systematica lly and comprehensively applied across multiple 86 \nunrelated parents across several different targets.  Muegge and colleagues have described a 87 \n“positional analog scanning” approach, which substitutes ligand non-polar hydrogens (C—Hs) 88 \nwith groups like CH ₃, OH,  Cl, F, and Br, and aromatic carbon atoms with nitrogen 21.  In 89 \nretrospective analyses of the ChEMBL24 database22,23, effects on affinity, functional activity, 90 \nsolubility, clearance, and permeability  were studied. Among over 110,000 matched molecular 91 \npairs (MMPs), about 30% of the time of one analog improved  over 3-fold in affinity /potency 92 \nversus the other.  Our understanding of this observation  is tempered by design and reporting  93 \nbiases toward compound improvement , by the variability in experimental conditions across 94 \ndatasets that can reduce reproducibility15, and by the few parent compounds that had multiple 95 \nconservative changes. Thus, it seemed interesting to create a set of parent ligands that were 96 \nsystematically and comprehensively modified by small perturbations, without design and 97 \nwithout a bias toward improvement, using the positional analog scanning approach 21,24 where 98 \nthe differences between parent and the suite of analogs were quantified by a single lab.  99 \n 100 \nTo create what we will refer to as  a random background set for ligand optimization, we 101 \nchose 18 parent ligan ds spanning six targets, including three GPCRs (alpha2A adrenergic 102 \nreceptor (2a), μ-opioid receptor (MOR) and cannabinoid receptor type 2  (CB2)), one 103 \ntransporter (the serotonin transporter, SERT) and two soluble enzymes (AmpC β-lactamase 104 \n(AmpC), and Macrodomain 1 of SARS-CoV-2 (Mac1)). This set was chosen for targets that we 105 \nhad under experimental control and is admittedly far from comprehensive: we explore only class 106 \nA GPCRs, no ion channels nor kinases are represented, only one transporter, and only two of 107 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 5 \nalmost innumerable soluble enzymes.  The 18 parent compounds were chosen for similar 108 \npragmatic reasons—they were molecules to which we had ready access.  Still, these proteins 109 \nprovide a range of target classes, while the 18 parents cover a wide range of physico-chemical 110 \nproperties (Extended Data Table 2) and values, ranging from 1 nM to 43 µM (Table 1), and 111 \nmost were topologically unrelated . For each of the 18 parents we systematically explored all 112 \nanalogs accessible by a single atom modification where such modifications were synthetically 113 \nfeasible—typically costing no more than $400 to acquire—and where they did not modify the 114 \nnet charge of the parent at physiological pH.   The single atom modifications involved 115 \nsubstituting carbon hydrogens with  one of methyl 25, hydroxyl 26, chloro 27, fluoro 28, or rarely 116 \nbromo, or substituting aromatic carbon atoms with nitrogen s (Figure 1).  These changes are 117 \nnot comprehensive, but they do represent changes preferred by medicinal chemists while 118 \navoiding those that would introduce oxidation liabilities (e.g., adding a thiol or aromatic sulfur) 119 \nor changes of valence (e.g., nitrogen to oxygen or sulfur) ; moreover, they were made without 120 \ndesign and as systematically as pragmatism would allow. 121 \n 122 \nThese substitutions were systema tically and comprehensively applied without design, 123 \nensuring an approach unbiased by anything other than synthetic accessibility and price (which 124 \nwe admit are meaningful constraints, as we make clear below) . Overall, 257 single-atom 125 \nanalogs were synthesize d and tested for changes in activity on the target, allowing us to 126 \nmeasure how often we might expect affinity/potency to improve >10-fold by minimal 127 \nperturbation.  This begins to provide a background expectation for how often a designed analog 128 \nmight meaningfully improve affinity/potency.  As we will show, these undesigned changes had 129 \nan unexpectedly high success rate, suggesting also a systematic strategy for optimization, even 130 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 6 \nthough that was not our primary goal.  Since affinity/potency maturation is only one criterion for 131 \nadvancing from hit to candidate, w e also measured in vitro pharmacokinetic properties for the 132 \nanalogs, including permeability, metabolic stability, plasma protein binding, and three other 133 \nterms, versus the parent compounds.  These pharmacokinetic studies explore what are perhaps 134 \ninevitable trade-offs between pharmacokinetics and target affinity/potency in ligand optimization.   135 \n 136 \nFig.1: Experiment and Analysis Workflow for One Heavy Atom Analogs.  a, Parent ligands from 137 \nsix drug targets (the aminergic GPCR Alpha2a-AR, the lipid GPCR CB2R, the peptide GPCR µOR, the 138 \nserotonin SLC transporter SERT, and the soluble enzymes AmpC and Mac1) were modified by one-at-139 \na-time, one atom substitutions: HCH3, HCl, HOH, HF, and ring CN. These modifications 140 \nwere made systematically and comprehensively; only compounds that were too expensive to 141 \nsynthesize, or that changed the charge state of the parent were left out.  The analogs were tested for 142 \nchanges in affinity/potency, solubility, plasma protein binding, plasma stability, hepatic microsomal 143 \nstability, permeability, and hERG inhibition, versus their parent ligands.  144 \n 145 \n 146 \nRESULTS 147 \nWe began with 18 parent ligands, most of which are unrelated to one another (ECFP4 148 \nTanimoto coefficients, Tcs, <0.35), with affinities ranging from 1 nM to 43 µM and molecular 149 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 7 \nweights (MW) ranging from 200 to 400 amu (Extended Data Figure 2). Each of the 18 parents 150 \nwas modified, one atom at a time, with five types of single atom substitutions (Figure 1a, central 151 \npanel), subject only to the expense and charge modification constraints (above).   About 7% of 152 \nthe possible molecules met these criteria  and of those ~80% were successfully synthesized, 153 \nresulting in 257 analogs. In the analyses that follow, we pool all 257 analogs for analysis, which 154 \nreduces the  noise from a single scaffold or single target  and improves statistical power 155 \n(Supplementary Figure 1). 156 \n 157 \nThese 257 analogs were tested for activity changes (Kd, IC50, or EC50) versus the parent 158 \nligands across the six targets.  For the μ-opioid receptor, EC50 values were determined in cAMP 159 \nGlo-sensor assays. 2A and CB2 ligands were assessed by cAMP assays and by radioligand 160 \ncompetition binding  with 3H-rauwolscine or [³H]CP-55940, respectively . Serotonin transporter 161 \ninteractions were analyzed using [3H]-citalopram binding assays. AmpC β -lactamase (AmpC) 162 \ninhibition was measured enzymatically while that of  Mac1 was measured by ligand 163 \ndisplacement. For every parent, and for most receptors overall, we only report changes for one 164 \ntype of activity, i.e., if a parent K i is reported than only K i values are reported for its analogs, 165 \nand if a parent is an agonist, only EC 50s are reported for its analogs.  Each functional or 166 \ninhibition assay is done in reference to a literature positive control molecule ( Supplementary 167 \nFigure 2). Representative concentration–response curves and the corresponding Z values for 168 \neach assay are prov ided in (Supplementary Figure 3). Especially for agonists, the effect of 169 \nreceptor expression is controlled for (Supplementary Figure 4, 5).  Because changes in 170 \nagonist efficacy can sometimes affect interpretation of EC 50, we inspect efficacy values to 171 \nensure that they are similar between parent and analog; full concentration-response curves and 172 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 8 \nanalyses are shown for every parent-analog pair, allowing for direct inspection of these results.    173 \n 174 \nOverall, 23% of the analogs  affected activity by <2-fold (fold changes rounded to the 175 \nnearest integer)  (Supplementary Table 1 , Supplementary Figure 1), while 30% reduced 176 \nactivity by >10-fold (Extended Data Figure 1 b).  Intriguingly, of the 257 analogs tested, 29, or 177 \n11.3%, improved activity by >10-fold rounding to the nearest integer (Table 1).  We note that 4 178 \nof these 29 had fold -changes of between 9.7 and 9.9 to  1 decimal place; we consider these 179 \neffectively 10 -fold improvements and used the same convention in analyzing the public 180 \nChEMBL data (see Table 1  and Supplementary Table S1  for details) .  Such >10-fold 181 \nimprovement analogs were found for five of the six targets, and for 10 the 18 parents (Table 1).  182 \nThe effects and chemical identities of all analogs relative to their parents are listed in 183 \nSupplementary Table 1. 184 \n 185 \nTable 1.  Analogs from all parents improving affinity/potency by >3x or >10x 186 \nProtein Assay \nreadout \nParent Ligand Parent \nAffinity/potency \nTotal \nAnalogs \nAnalogs with ≥3X \nAffinity/potency \nimprovement \nAnalogs with ≥10X \nAffinity/potency \nimprovement \nMOR EC50 Z4407498716 9.2 nM 16 2 2 \nalpha2 EC50 Z2750653629 1.4 nM 31 7 4a \nalpha2 Ki Z3034773248 22.1 µM 18 3 1 \nalpha2 EC50 Z4376630014 8.3 nM 10 2 0 \nalpha2 Ki Z8727395870 13 µM 18 1 0 \nAmpC Ki Z2610488449 24.2 µM 10 2 0 \ncb2 Ki Z52076138 2.3 µM 7 7 6b \ncb2 Ki Z6969215903 0.4 µM 6 0 0 \ncb2 EC50 Z8184698918 0.1 µM 20 2 2 \nMac1 IC50 Z5398393122 2.2 µM 5 0 0 \nMac1 IC50 Z1039063794 0.7 µM 12 0 0 \nMac1 IC50 Z7534253453 8.9 µM 22 8 2 \nSERT Ki Z2009978218 1.1 µM 14 4 0 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 9 \nSERT Ki Z2573292480 0.08 µM 11 1 0 \nSERT Ki Z2573292509 0.2 µM 9 2 1 \nSERT Ki Z6971277399 13.4 µM 15 10 4 \nSERT Ki Z8727393896 42.7 µM 14 8 3c \nSERT Ki Z8731642686 1.8 µM 19 10 4 \nTotal  All Parents  257 69 29 \n*Fold-changes are displayed as nearest -integer values; exact (unrounded) fold -changes are provided in 187 \nSupplementary Table S1. Using exact fold -changes, 62 analogs improve ≥3-fold and 25 improve ≥10-188 \nfold. a. One of these had a fold change of 9.7.  b. Two of these had fold-changes of 9.7 and 9.8.  c. One 189 \nof these had a fold-change of 9.9, see Supplementary Table S1. 190 \n 191 \nWe were interested in which perturbations had the biggest effects, how the rates of 192 \naffinity/potency improvements compared to the literature, with its admitted biases, and if 193 \nphysical properties of the parent compounds or of the binding sites correlated with greater 194 \nlikelihoods of affinity/potency improvements.  Among the analogs, methyl substitutions were the 195 \nmost likely to improve activity >10-fold, with chlorine substitutions the second most likely; for >3-196 \nfold improvements, of which there were 69 among the 257 analogs, both chlorine and methyl 197 \nsubstitutions were the most common (Figure 2a ).  Given the similar size and similar 198 \nhydrophobicity of methyls and chlorines, their similar effects may be rationalized. Conversely, 199 \nfluorine substitutions, which are often used to increas e metabolic stability, yielded no >10-fold 200 \nand few >3-fold improvements in activity.  Aromatic carbon to nitrogen substitutions also rarely 201 \nimproved affinity/potency, though as we will see they often had the most favorable impacts on 202 \npharmacokinetic properties.   203 \n 204 \nBecause increases in hydrophobicity will increase affinity/potency simply by disfavoring 205 \nthe unbound state, and have physical property liabilities, it is us eful to consider the improved 206 \naffinity/potency of the analogs in light of ligand efficiency (LE) and liphophilic efficiency (LiPe) , 207 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 10 \nwhich control for these effects .  If affinity/potency improvements were largely driven by 208 \nhydrophobicity we would expect to see both terms deteriorate for the analogs.  Instead, for 209 \nanalogs that improved >10-fold both ligand efficiency and lipophilic efficiency much improved.  210 \nFor instance, an analog with an added methyl, chloro or other atom that improved in binding or 211 \nin EC50 by even 3-fold enjoyed a ligand efficiency of 0.65 kcal/atom for the added atom, well 212 \nabove the 0.3 kcal/atom that is thought to be useful for drug optimization 29,30. Meanwhile, 213 \nanalogs that improved activity >10-fold had ligand efficiency of 1.4 for a single added atom (Ext 214 \nData Figure 4 ), close to the efficiency limit 31 (several even exceeded single atom ligand 215 \nefficiencies of 2.0).  Similarly, every analog that improved activity by >10-fold also improved 216 \nlipophilic efficiency  (Ext. Data Figure  4).  More broadly, while there was little correlation 217 \nbetween clopgP and pKi or logEC50 over the the 257 analogs (Ext Data Figure 4b), there 218 \nwas a strong linear correlation between both ligand efficiency and activity change and between 219 \nlipophilic ligand efficiency and activity change —in both cases, substantially improved activity 220 \nwas accompanied by improved ligand and lipophilic efficiency.  For analogs with much improved 221 \nactivity—especially those improved by >10 -fold but also for those improved by >3 -fold—the 222 \nimprovements cannot be simply laid at the door of increased hydrophobicity.   223 \n 224 \nWe investigated whether  there were correlations between likelihood of potency  225 \nimprovement and ligand and binding site properties.  One might expect that certain ligands or 226 \nsites might better lend themselves to affinity/potency improvement, which would affect the 227 \ndomain of applicability of this study . For instance, weaker ligands might be easier to optimize, 228 \nor smaller binding sites might be better suited to big affinity/potency jumps.  We calculated if 229 \nthe parent molecular weight, binding site volume, the ligandability of the binding sites 230 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 11 \n(calculated using Dscore 32), or the affinity/potency of the starting parent compound was 231 \ncorrelated with the likelihood of finding analogs with substantial affinity/potency improvement. 232 \nUsing the fraction of analogs achieving ≥10 -fold improvement as the outcome, none of these 233 \nproperties showed a statistically meaningful correlation  (Figure 2b ).  If one considers the 234 \nmagnitude of the affinity/potency change overall, and not just whether it improved >10-fold, 235 \ncorrelations did emerge, with weaker vs stronger parents more likely to support improved affinity 236 \n(R = −0.34, p = 3.3×10 −7) and with larger bi nding sites also tending to do so (R = 0.23, p = 237 \n4.3×10−4) (parent size and site ligandability  continued to show no association with affinity 238 \nimprovement).  While these trends were echoed among matched pairs in ChEMBL (Extended 239 \nData Figure 1), the correlations overall were weak (even an R of -0.34 explains only 10% of 240 \nset variance), and as activity changes drop below a 3-fold effect our confidence in the 241 \nexperimental values begins to drop.   Taken together,  these results suggest that : (1)  as a 242 \nstrategy, systematic small perturbations can be unexpectedly successful at improving ligand 243 \naffinity/potency21,24, perhaps even compared to modern design -heavy approaches33; (2) this 244 \neffect is common across diverse parents and binding sites, (3) finds echo in the literature, and 245 \n(4) random expectations for substantial affinity/potency improvement may be as high as 11% 246 \n(95% CI 7.5%-15.2%).   247 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 12 \n  248 \nFig.2 Property correlations of  affinity/potency improvement and comparison with literature 249 \nvalues. a. The fraction of compounds improving by >3- or >10-fold in activity by type of single 250 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 13 \nmodification (CH₃, F, Cl, Br, N, OH).  b. Correlation heatmap between physicochemical properties of 251 \nbinding pockets and parent compounds with log₁₀ of affinity/potency fold-change. c. The fraction of 252 \nsingle atom  modifications (CH ₃, F, Cl, N, OH)  that improve a ctivity, comparing the experimental 253 \nobservations in this study with results from the ChEMBL database.  Shaded regions indicate 95% 254 \nconfidence intervals, estimated from 10,000 iterations of bootstrap resampling .  The ChEMBL ligands 255 \nare a composite of 54,021 parents and 79,984 small change analogs; for only 4.7% of these were there 256 \nmore than three analogs per parent.  257 \n 258 \nTo investigate the structural bases for the affinity/potency changes, we determined the 259 \nstructure of the parent inhibitors in complex with Mac1  and six analogs from the '3122, 12 260 \nanalogs from the '3453 series, and four analogs from the ‘3794 series. These structures were 261 \ndetermined to between 0.97 and 1.02  Å resolution (Supplementary Table 2 ), supporting 262 \natomic resolution analysis. The analogs with structures determined represented all six types of 263 \nsingle-atom modifications in this study; 15 of them differed in affinity/potency from their parents 264 \nby >3x, two by over 10-fold, and among themselves by up to 680-fold.  265 \n 266 \nFor some parents, analogs superposed well facilitating analysis of the effects.  This was 267 \noften true for the Mac1 ‘3453 (IC50 8.9 µM) inhibitor series, for instance (Figure 3a). Displayed 268 \nat the ortho-position of the phenyl ring the chloro of ‘0676 (IC50 21 µM) projects into solvent and 269 \nloses activity versus the parent ( Figure 3b).  Moving this chloro to the meta -position ('1304), 270 \nwhere it packs with Phe132, Ile131, and Gly48 , improves inhibition 11 -fold (p-value <0.01) to 271 \n1.9 µM (Figure 3c).  Moving this substitution one atom further over, to the para-position, had 272 \nlittle effect ('6343, IC50 2.1 µM) though a methyl at the same position ('9870, IC50 7.6 µM) loses 273 \n4-fold affinity (p-value <0.01), presumably reflecting its poorer packing versus the chloro analog.   274 \n 275 \nIf many of the effects of the small perturbation analogs could be explained post hoc from 276 \ntheir structures, however, fewer were easily anticipated.  Even the apparent simplicity of the 277 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 14 \n‘3453 series is belied by structural accommodations in the enzyme site.  For instance, the para-278 \nchloro of the 2µM ‘6343 would have clashed with the enzyme without Phe132 and Ile131 279 \nrotating away.  In doing so, Phe132 adopts a partially eclipsed conformation. Presumably the 280 \nresulting strain is overcome by the improved packing of the buried para-chloro, but the outcome 281 \nof this balance of terms seems difficult to intuit (encouragingly, free energy calculations did 282 \ncorrectly predict that the para -and meta-chloro ‘6343 and ‘1304, respectively, would improve 283 \naffinity while the ortho -chloro ‘0676 would lose it, though the agreement was more qualitative 284 \nthan quantitative; see below). Conversely, the methyl analog (‘9249 , 53 µM) of '0676 (IC50 22 285 \nµM) is buried from solvent, not exposed, and might be expected to be more potent than its 286 \nchloro cousin, but instead it loses another 2.5-fold in Ki (p-value <0.01; here, the opposite effect 287 \nwas anticipated by the free energy calculation, below).  Meanwhile, in the complex between 288 \n‘4905 and the 2a adrenergic receptor, an added methyl group fit s into a pre -existing sub-289 \npocket and qualitatively its improved potency is reasonable. Quantitatively, its 52-fold effect on 290 \nEC50 is at the outer edge of what might be expected by ligand -efficiency, especially since the 291 \nmethyl buries a serine hydroxyl.  This is echoed by the free energy calculation, which here too 292 \ncorrectly predicts improved activity, but only by 2 - not 52 -fold.  Th ese are examples of 293 \ncomplexes where the analogs largely superpose on the parent and each other.  In complexes 294 \nwhere the single-atom changes led to substantial inhibitor movement , or to the adoption of 295 \nmultiple ligand conformations, prediction was harder still (Extended Data Fig 5). For example, 296 \nthe Mac1 methyl analog ‘3176 repositions the inhibitors in the pocket (Extended Data Fig. 5b), 297 \nwhile compounds ‘9249 and ‘3194 adopt two distinct conformations in the site (Extended Data 298 \nFig. 5b), a phenomenon that i s likely underreported in the PDB 34 but is readily seen at ultra -299 \nhigh resolution.   Taken together, this set of perturbations may provide an interesting and 300 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 15 \nsometimes challenging test for prediction methods, and su ggests that there may be merit in 301 \nsystematic small perturbations as a strategy, though showing that was not a goal of this study.   302 \n 303 \nFigure 3. Structural basis of potency changes induced by single -atom modifications for Mac1 304 \ninhibitors.  a. Chemical structure of the ’3453 parent and overlay of X -ray crystal structures of 12 305 \nanalogs. b.  Alignment of ‘0676 (21 μM) and ‘9249 (53 μM) with the ‘3453 parent showing the difference 306 \nin binding pose for the ortho -chloro and ortho-methyl substituted analogs. c. Comparison of ‘6343 (2.1 307 \nμM), ‘1304 (1.9 μM) and ‘0676 (21 μM) reveals that repositioning the aryl chloride disrupts packing with 308 \nI131 and F132, reducing potency. d. Para-methyl (‘9870, 7.6 μM), para -chloro (‘6343, 2.1 μM) 309 \nsubstitutions have different effects despite similar sterics. Fluoro analog ‘6404 (5.5 μM) shows weaker 310 \nbinding than the ‘chloro (‘6343), likely reflecting weaker non -covalent interactions. e. Nitrogen 311 \nsubstitutions (C→N) in ‘3169, ‘6627, and ‘8675 all pay increased desolvation penalties. Compared to 312 \ntheir shared parent (8.8 μM), ‘8675 (meta -N) improved potency nearly 2-fold (4.5 μM), ‘6627 (8.7 μM) 313 \nshowed little improvement, and ‘3169 (10 μM) is slightly worsened. 314 \n 315 \nIn optimizing chemical probes and leads , compound pharmacokinetics (PK) is as 316 \nimportant as molecular target affinity/potency 35,36. We thus explored how the small perturbation 317 \nanalogs affected the PK properties versus those of the parent ligands, and how these changes 318 \nrelated alterations of affinity/potency.  We measured the following in vitro PK properties: 319 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 16 \nmetabolic stability in microsomes  (half-life), plasma stability (half-life), plasma protein binding 320 \n(fraction unbound), solubility  (uM), hERG inhibition (IC50, uM) , and membrane permeability 321 \n(PAMPA assay) (cm/s) of both the parent ligands and their small change analogs (Figure 4, 322 \nSupplementary Table 3-8).  While these are all in vitro measurements, they are proxies that 323 \nare considered predictive of in vivo PK37-39.  Some of  the small change analogs showed 324 \nmeaningful improvements relative to the parent ligands for each of these properties (Extended 325 \nData Table 2).  For example, 12.4% of the analogs had ≥3-fold improvement in microsomal 326 \nstability, while 20.8% of them had a >3-fold improvement in plasma stability, versus the parents. 327 \nFor 13.8% of the analogs, solubility improved by ≥3-fold, while 6.2% of them improved >3-fold 328 \nin permeability.  Examining the fraction unbound in plasma, 13.2% of the analogs improved ≥3-329 \nfold. These findings illustrate that each property is amenable to meaningful optimization even 330 \nby simple changes. We also observed a substantial fraction of analogs with ≥3 -fold decreases 331 \nin these same properties (Extended Data Figure 6).  Considering what are perhaps the three 332 \nmost impactful in vitro PK measurements, microsomal stability, fraction unbound, and 333 \npermeability, 36% of the analogs suffered >3 -fold losses in at least one of these , almost all 334 \n(96%) suffered at least some deterioration by the same criterion, and 37% deteriorated in all 335 \nthree in vitro PK properties. 336 \n 337 \nIf many of the small perturbation analogs improved substantially in individual PK 338 \nproperties, none of those with >10-fold increased potency improved or even maintained more 339 \nthan one of what we might consider the three most important ones: metabolic stability, PAMPA 340 \npermeability, and fraction unbound in the plasma (Figure 4.a, Extended Data Figure 7); most 341 \nwere worse by all three properties. Indeed, the changes in in vitro PK were, at best, orthogonal 342 \nto affinity/potency fold-change (Figure 4b), and several properties were anti-correlated with it.  343 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 17 \nThis is born out on an atom -by-atom level: those atoms that most often improved 344 \naffinity/potency substantially, like methyls and chloros (Figure 2a), were least associated with 345 \nimproved pharmacokinetic properties, such as microsomal stability or fraction unbound (Figure 346 \n4c, Extended Data Figure 8).  Meanwhile, those atoms least likely to improve affinity/potency 347 \nsubstantially, like aromatic C N, were most likely to improve microsomal stability or fraction 348 \nunbound. Density and cumulative distribution plots illustrate this divergence in effect (Extended 349 \nData Figure 8), with distinct shifts in property distributions across modifications. More broadly, 350 \nmost of the pharmacokinetic properties were largely independent of one another, with no 351 \ncorrelations exceeding |ρ| > 0.5 and those that were statistically significant having Spearman ρ 352 \nvalues ranging only from |0.16| to |0.38|.  What emerges is a frustrated landscape for ligand 353 \noptimization, with affinity/potency improvements counterbalanced by often worsening 354 \npharmacokinetics.  While this trade-off has been noted in medicinal chemistry, 40,41 this set 355 \nquantifies its systematic impact.  356 \n  357 \nFig.4 A frustrated landscape for ligand optimization.  a, Fold-changes in pharmacokinetic 358 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 18 \nproperties for the 29 analogs that were >10-fold improved in affinity/potency versus a parent. b, 359 \nCorrelations between affinity/potency fold changes and those of the other PK properties. c, The 360 \nfraction of single atom changes that were >3x fold improved in each parameter. Error bars represent 361 \n95% confidence intervals, estimated from 10,000 iterations of bootstrap resampling. 362 \n 363 \nThe current benchmark only comprises several hundred analogs.  If we could predict the 364 \nresults of the small perturbations, rather than having to experimentally test them, the set could 365 \nbe much extended.  We therefore asked how well modern methods could predict the relative  366 \naffinity/potency and pharmacokinetic changes we observed.  In blinded experiments, we used 367 \nfree energy perturbation , arguably the highest level of theory available to the field, with the 368 \nprogram FEP+ (Schrodinger, New York) to predict the changes in activity between the parents 369 \nand analogs (Figure 5 a-c).  Similarly, we asked how well we could predict in vitro  370 \npharmacokinetic changes between the parents and analogs, using widely accessible tools  371 \n(Figure 5d).   372 \n 373 \nThere was a strong overall correlation between the FEP+-predicted and experimentally 374 \nmeasured binding free energies (ΔG), with most predictions (190) falling within ±2 kcal/mol of 375 \nthe experimental values, and 131 within ±1 kcal/mol, while the set had a mean unsigned error 376 \n(MUE) to ΔG as 1.07 kcal/mol (Figure 5a,b). Of the 18 parents, the FEP+ predictions on 14 had 377 \nΔG MUE below 1.2 kcal/mol (Supplementary Table 9). If one considers the correlations 378 \nqualitatively, of the 35 analogs predicted to lose affinity by 1 kcal/mol or more by FEP+ (i.e., in 379 \nthe range of confident prediction), 20 did lose affinity/potency by experiment.  Meanwhile, of the 380 \n34 predicted to increase in affinity/potency by more than 1 kcal/mol by FEP +, 14 did so  381 \nexperimentally while 12 were experimentally neutral and 8 instead showed decreased potency. 382 \nCategorization with a 1  kcal/mol threshold reveals a significant association between the 383 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 19 \ndirection and strength of FEP+ predictions and experimentally observed affinity/potency 384 \nchanges (² of 36.4 and a p-value < 0.001) (Supplementary Table10). 385 \n 386 \nIt is useful to note several of the challenges facing accurate FEP  predictions for this 387 \nseries, which future work may share.  The predictive power was likely weakened by our inability 388 \nto measure affinity/potency for weak analogs, diminishing the overall range of affinities and 389 \nmaking it impossible to measure affinities predicted to be very low by FEP +.  We used an 390 \nautomated workflow to generate input poses for the calculations; on visual inspection, most of 391 \nthe outliers were caused  by suboptimal initial poses of analogs.   The accuracy of FEP 392 \ncalculations is strongly impacted by the quality of the ligand -complexed structures42, and the 393 \nFEP+ predictions here began with docking poses, which are less reliable than experimental 394 \nstructures. For the 25 analog structures that we determine d by crystallography (for Mac1 395 \nanalogs) and by cryoEM (for an 2a analog, below), there were  many cases where the 396 \npredicted and experimental poses closely superimposed, and for these the FEP+ energies were 397 \nlargely consistent with the experimental measurements (Figure 5C).  398 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 20 \n 399 \nFigure 5: Accuracy of computational predictions of analog relative affinity and in vitro 400 \npharmacokinetics.  a. Comparison of FEP-predicted and experimentally measured binding free 401 \nenergies (ΔG). Dashed lines indicate ±1 and ±2 kcal/mol error margins.  b. Correlation between 402 \nFEP-predicted and experimental measured relative binding free energy changes (ΔΔG). c. An 403 \nexample of good structural alignment between the predicted (cyan) and experimental (grey) 404 \nposes, accompanied by a small error in the predicted ΔΔG.  d. Heatmap of Spearman 405 \ncorrelations between AI-predicted and experimentally measured fold changes across six in vitro 406 \nPK properties, evaluated using three models. Color intensity indicates correlation strength; 407 \nsignificance: p < 0.05 (*), < 0.01 (**), < 0.001 (***). 408 \n 409 \nRecently, AI-based platforms have been developed to facilitate ligand optimization by 410 \npredicting ADMET properties with increasing accuracy and scale, guiding experiment43,44. We 411 \ncompared the ADMET predictions from ADMET -AI, ADMETLab 3.0, and Deep -PK45-47 to the 412 \nexperimentally measured values. The predictions of all three were significantly correlated with 413 \nexperimental measurements of plasma fraction unbound and with permeability, while 414 \nmicrosomal stability , plasma stability  and solubility, which are widely considered difficult to 415 \npredict48,49, were essentially uncorrelated.  Correlations in the 0. 4 to 0.5 range, as they often 416 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 21 \nwere for fraction unbound and permeability, are strong enough to help guide compound 417 \nselection in many circumstances, and these open -access tools may have broad impact.  Still, 418 \neven with a correlation of 0. 53 in fraction unbound, the highest we observed, 28 analogs that 419 \nhad substantially increased fraction unbound experimentally were predicted to have lower 420 \nlevels than their parents, and  16 analogs that had higher protein binding than their parents 421 \nexperimentally were predicted to have less.  Taken together, while the FEP and ADMET 422 \npredictions can profitably model the effects of the small perturbation analogs, their correlations 423 \nwith experiment are not yet strong enough to allow this set to be extended by calculation alone.   424 \n 425 \nMulti-parameter optimization for an improved analgesic. Notwithstanding the 426 \nfrustrated landscape for ligand affinity /potency and pharmacokinetic improvement, medicinal 427 \nchemists can often navigate the multi-parameter optimization to overall improve the properties 428 \nof lead molecules.  In this study, too, there were analogs that improved in affinity/potency with 429 \nonly modest sacrifice in pharmacokinetics (PK).  Among these was an analog of an 2a parent, 430 \ncompound ‘4905, where a single methyl group substitution improved potency 52-fold (Figure 431 \n6a, b).  To understand this improvement at atomic  resolution, we determined the structure of 432 \n‘4905/alpha2a/G protein complex by single particle cryoEM to 2.8 Å resolution.  Encouragingly, 433 \nthe experimental structure of ‘4905 superposed closely with the docking prediction for the 434 \nparent ‘3629 and with the experimental structure of a previously described lead member of this 435 \nfamily of alpha2a agonists, ‘ 9087, suggesting that the methyl perturbation did not change the 436 \norientation of the analog, and that this family may be understood in the same structural context. 437 \nFrom the cAMP dose –response curves, we obser ved a slight upward deflection at high concentrations 438 \nfor ‘4905. Consistent with this, PTX -treated cAMP assays showed a residual, concentration -dependent 439 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 22 \nincrease in cAMP for ‘4905 but not 3629 (Supplementary Figure 6), suggesting secondary Gs coupling 440 \nby the ligand. Structurally, the extra methyl in 4905 fills a hydrophobic subpocket and shifts Y409 6.55 441 \n(implicated in Gs activation), which may bias the receptor toward conformations permissive for Gs 442 \nengagement. Improved steric complementarity with residues  in TM5 and TM3, and a closer 443 \ndistance between the partly cationic bridging nitrogen with the crucial Asp128 on TM3, helps to 444 \nexplain the improved potency of ‘4905 over its parent.   445 \n 446 \nWhile the improved potency of ‘4905 came at a cost to its in vitro pharmacokinetics, the 447 \neffects were relatively modest: Compared to its parent, ‘3629, the fraction unbound decreased 448 \nby 57% and the microsomal half -life was reduced by 41%.  (Supplementary Table 3-8).  The 449 \nin vivo PK of ‘4905 qualitatively reflected the in vitro results, with CSF concentrations—a proxy 450 \nfor fraction unbound in the brain50—reduced by about 36%, while the half-life was reduced from 451 \n196 to 15.2 minutes (Supplementary Table 3-8).  These size of these effects suggested that 452 \nthey might be outweighed by the analog’s great increase in activity.  Accordingly, we compared 453 \nthe in vivo  analgesia conferred  by ‘4905 to its parent ‘3629 and to PS75, the most potent 454 \nanalgesic known for this family  of alpha2a agonists 51.  Despite the deterioration in 455 \npharmacokinetic properties, in mouse nociception assays including tail flick, neuropathic pain 456 \n(SNI), and hot-plate the new analog ‘ 4905 was far more potent than the parent ‘3629 and at 457 \nleast as potent as PS75 (Figure 6 c-e).  If ligand optimization is a frustrated landscape, it 458 \nremains a navigable one as is well-known to medicinal chemistry\n41\n. In addition to beginning to 459 \nprovide a random background for such efforts, this study supports a strategy by which that 460 \nlandscape may be reconnoitered21,24.    461 \n 462 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 23 \n   463 \nFigure 6. A single-atom addition to an alpha2a receptor agonist creates a more potent 464 \nanalgesic in vivo.  a. Docking pose of parent ‘3629 superposed on the cryoEM structure of analog 465 \n‘4905, which differs from ‘3629 by the addition of a single methyl group.    466 \nb. ‘4905 is 50-fold more potent than the parent ‘3629 in receptor activation in vitro.  We note that the 467 \napparent reduced efficacy through Gi coupling at concentrations greater than 1 nM likely reflects the 468 \nactivation of Gs.  c-g.  ‘4905 is more potent than the parent ‘3629 in c-e. Reflex pain as measured in 469 \ntail flick response to heat; f. Reversal of mechanical allodynia in a neuropathic pain model (SNI). g. 470 \nEfficacy in reducing thermal acute pain assessed via the hotplate test. 471 \n472 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 24 \nDISCUSSION 473 \nWhile ligand optimization is at the heart of drug discovery, quantifying the success of 474 \nstrategies to modify ligands has been difficult owing to a lack of a random background  475 \ncomparison set.  Four key observations emerge from this study.  First, systematic and unbiased 476 \nsmall perturbations to 18 different scaffolds revealed that 11.2% of the analogs improved >10-477 \nfold in affinity /potency versus the parent (Table 1); “magic methyls” 25,52 were common in this 478 \nset.  These improvements were observed against five of six receptors targeted, across physical 479 \nproperties of target and ligand, and across five orders of magnitude of parent  affinity/potency.  480 \nThis begins to establish a background expectation for affinity /potency optimization against 481 \nwhich methods and design s may be compared .  It may also suggest a simple, unbiased 482 \napproach for achieving substantial jumps in affinity/potency, improvements that will not always 483 \nbe obvious from structural analyses or even high -level simulations (Figures 2 and 5) but may 484 \ncomplement them.  Second, improved affinity/potency typically came at the cost of ligand 485 \npharmacokinetics (PK), with plasma fraction unbound, stability to liver microsomes, and 486 \npermeability—among other terms —declining as analog affinity /potency increased.  The 487 \nindividual in vitro pharmacokinetic terms were orthogonal to each other, and several were anti-488 \ncorrelated with affinity/potency improvement (Figure 4).  This suggests a frustrated landscape 489 \nfor ligand optimization.  Although the qualitative challenges and trade-offs in ligand optimization 490 \nare known to medicinal chemists 6,40,41, this study begins to quantify them at scale and across 491 \na range  of target classes and ligand properties .  Third, computational prediction of 492 \naffinity/potency and PK had some success anticipating the trends we observed ( Figure 5) 493 \nimplying that they may even tually guide compound selection .  Still, their correlations with 494 \nexperiment remained loose enough to preclude replacing systematic exploration and testing of 495 \nmolecules as this background set is further expanded , something seen with other potency 496 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 25 \nprediction methods53.  Fourth, while potency trades-off against PK, it is possible to find analogs 497 \nwhere it rises sufficiently to overcome drops in exposure and half-life.  A testament to this is the 498 \nin vivo potency of the 2a analog ‘4905, which despite suffering these trade-offs nevertheless 499 \nis the most potent analgesic in vivo in a large series, and is certainly far more potent in vivo 500 \nthan its parent.  It may be interesting to model the implications of the orthogonal and anti -501 \ncorrelated terms here to understand how best to navigate this multi -parameter optimization 502 \nproblem.   503 \n 504 \nAt first glance, this study is reminiscent of  learnings from classical medicinal chemistry 505 \nwhere conservative and ligand-based approaches like Topliss Trees9 dominated the field.  But 506 \nthis would be to misremember the classic age of medicinal chemistry on two counts.  First, the 507 \nsystematic small perturbations explored here were rarely used then, partly because the building 508 \nblocks and reactions to support them were lacking.  The ability to systematically explore small 509 \nperturbations, across starting scaffolds and targets, reflects the advances in synthetic chemistry 510 \nover the last 50 years and also the new building-block approaches that have expanded readily 511 \navailable compounds 54,55.  Second,  the logic behind Topliss Trees  and other quantitative 512 \nstructure-activity approaches9,10 were anchored in ligand physical chemistry without reference 513 \nto receptor structure.  For instance, in Topliss Trees the success of methyl and chloro 514 \nderivatives were thought to teach opposite lessons because of their different el ectronic effects 515 \non the ligand, whereas the steric and non-polar properties of the two groups will seem similar 516 \nfrom the view “inside the receptor”52,56.Indeed, this is what we find in the similar affinity/potency 517 \neffects of the two substitutions.  Meanwhile, the advantage of a random background, so integral 518 \nto molecular biology, would have seemed foreign to classical medicinal chemistry. Thus, while 519 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 26 \na strategy of systematic small perturbations may smack of the “methyl, ethyl, propyl, butyl, futile” 520 \napproach with which classic medicinal chemistry has been tarred 57,58, it has a different basis in 521 \ntheory and reflects new synthetic opportunities.   522 \n 523 \nCertain limitations of this study  merit airing.  At 257 analogs and six targets, our set 524 \nremains small and inevitably, if unintentionally, biased.  We do not pretend that three class A 525 \nGPCRs, one transporter, and two soluble enzymes adequately represent pharmacologically 526 \nrelevant targets, nor that the 18 parent ligands can adequately represent ligand space.  527 \nExpanding the analysis to additional targets and ligand classes  might reveal different trends.  528 \nWe note that the six targets do recognize a range of chemotypes, including large polar co -529 \nfactors (Mac1) , large hydrophobic  lipids (CB2), anionic -lactams ( AmpC), cationic 530 \nneurotransmitters (2a and SERT), and peptides (µOR).  Experimentally, we assessed activity 531 \nchanges using radioligand competition binding assays (SERT, 2a), second messenger assays 532 \n(2a, CB2, and MOR), and binding and enzyme activity (Mac1 and AmpC) assays.  While ligand 533 \ncompetition correlates with Ki, agonist EC50s are affected by receptor expression.  Because we 534 \ncompared the relative activities of analogs versus their parents , controlled for receptor 535 \nexpression (Supplementary Figure 4), and measured effects in a low expression level domain 536 \n(Supplementary Figure 5), the effects of receptor expression on relative activity should be 537 \nmodest, though they cannot be completely discounted.  For these and related reasons, changes 538 \nof EC50 between agonist parent and analogs cannot be read as changes in affinity the way that 539 \nchanges in Ki can be, though EC 50 changes remain the relevant metric for agonist activity.  540 \nOverall, the impact of these effects may be inspected case by case in full co ncentration-541 \nresponse for any parent ligand pair with >3 -fold improvement in activity (Supplementary 542 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 27 \nFigure 2).  In the area of in vitro pharmacokinetics, our  use of liver microsomes rather than 543 \nhepatocytes to measure metabolic stability meant that some types of metabolism were missed, 544 \nincluding glucuronidation.  This might make groups like phenolic hydroxyls seem less labile 545 \nthan they would be in vivo.  More broadly,  for all but two molecules we only measured in vitro 546 \nnot in vivo pharmacokinetics.  Whereas the in vitro measurements qualitatively anticipated the 547 \nin vivo effects where we did measure them, and are widely used in ligand optimization, their 548 \nquantitative prediction of in vivo behavior is only approximate. 549 \n 550 \nThese limitations should not obs cure the main observations of this study.  Over 11.2% 551 \nof systematic and unbiased small perturbations improved  analog affinity/potency >10-fold, 552 \nbeginning to establish a background expectation for the likelihood of substantial ligand 553 \naffinity/potency improvement and a systematic approach to doing so 21,24.  Balancing this was 554 \na concomitant deterioration in ligand pharmacokinetics, which will lower the exposure and half-555 \nlife of a ligand in vivo, counteracting the improvements in affinity/potency.  Navigating this multi-556 \nparameter space is at the heart of medicinal chemistry; this study supports to the development 557 \nof quantitative models to do so.    558 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 28 \nReferences 559 \n1 Hughes, J. P., Rees, S., Kalindjian, S. B. & Philpott, K. L. Principles of early drug discovery. 560 \nBrit J Pharmacol 162, 1239-1249, doi:10.1111/j.1476-5381.2010.01127.x (2011). 561 \n2 Bleicher, K. H., Bohm, H. J., Muller, K. & Alanine, A. I. Hit and lead generation: beyond 562 \nhigh-throughput screening. Nat Rev Drug Discov 2, 369-378, doi:10.1038/nrd1086 (2003). 563 \n3 Paul, S. M.  et al. How to improve R&D productivity: the pharmaceutical industry's grand 564 \nchallenge. Nature Reviews Drug Discovery 9, 203-214, doi:10.1038/nrd3078 (2010). 565 \n4 Lipinski, C. A. Lead - and drug -like compounds: the rule -of-five revolution. Drug Disc ov 566 \nToday Technol 1, 337-341, doi:10.1016/j.ddtec.2004.11.007 (2004). 567 \n5 Stumpfe, D., Hu, H. & Bajorath, J. Evolving Concept of Activity Cliffs. ACS Omega 4, 14360-568 \n14368, doi:10.1021/acsomega.9b02221 (2019). 569 \n6 Leeson, P. D. & Springthorpe, B. The influence o f drug-like concepts on decision-making 570 \nin medicinal chemistry. Nat Rev Drug Discov 6, 881-890, doi:10.1038/nrd2445 (2007). 571 \n7 Wunberg, T. et al. Improving the hit-to-lead process: data-driven assessment of drug -like 572 \nand lead -like screening hits. Drug Disco v Today  11, 175 -180, doi:10.1016/S1359 -573 \n6446(05)03700-1 (2006). 574 \n8 Jorgensen, W. L. Efficient drug lead discovery and optimization. Acc Chem Res 42, 724-575 \n733, doi:10.1021/ar800236t (2009). 576 \n9 Topliss, J. G. Utilization of Operational Schemes for Analog Synthes is in Drug Design. J 577 \nMed Chem 15, 1006-&, doi:DOI 10.1021/jm00280a002 (1972). 578 \n10 Hansch, C. Citation Classic - Rho-Sigma-Pi-Analysis - a Method for the Correlation of 579 \nBiological-Activity and Chemical-Structure. Cc/Life Sci, 18-18 (1982). 580 \n11 Dong, J.  et al.  ADMETlab: a platform for systematic ADMET evaluation based on a 581 \ncomprehensively collected ADMET database. J Cheminformatics  10, doi:ARTN 29  582 \n10.1186/s13321-018-0283-x (2018). 583 \n12 Wang, L. et al. Accurate and Reliable Prediction of Relative Ligand Binding Po tency in 584 \nProspective Drug Discovery by Way of a Modern Free -Energy Calculation Protocol and 585 \nForce Field. J Am Chem Soc 137, 2695-2703, doi:10.1021/ja512751q (2015). 586 \n13 Klimovich, P. V., Shirts, M. R. & Mobley, D. L. Guidelines for the analysis of free energy 587 \ncalculations. J Comput Aided Mol Des  29, 397 -411, doi:10.1007/s10822 -015-9840-9 588 \n(2015). 589 \n14 Hajduk, P. J. & Sauer, D. R. Statistical analysis of the effects of common chemical 590 \nsubstituents on ligand potency. J Med Chem 51, 553-564, doi:10.1021/jm070838y (2008). 591 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 29 \n15 Kramer, C., Kalliokoski, T., Gedeck, P. & Vulpetti, A. The Experimental Uncertainty  of 592 \nHeterogeneous Public Data. J Med Chem 55, 5165-5173, doi:10.1021/jm300131x (2012). 593 \n16 Booker, T. R., Jackson, B. C. & Keightley, P. D. Detecting positive selection in the genome. 594 \nBmc Biol 15, doi:ARTN 98 10.1186/s12915-017-0434-y (2017). 595 \n17 Pritchard, J. K. & Cox, N. J. The allelic architecture of human disease genes: common 596 \ndisease - common variant ... or not? Hum Mol Genet  11, 2417 -2423, doi:DOI 597 \n10.1093/hmg/11.20.2417 (2002). 598 \n18 Manolio, T. A. et al. Finding the missing heritability of complex diseases. Nature 461, 747-599 \n753, doi:10.1038/nature08494 (2009). 600 \n19 Fowler, D. M. & Fields, S. Deep mutational scanning: a new style of protein science. Nat 601 \nMethods 11, 801-807, doi:10.1038/Nmeth.3027 (2014). 602 \n20 Wells, J. A. Systematic Mutational Analyses of Protein Protein Interfaces. Method Enzymol 603 \n202, 390-411 (1991). 604 \n21 Pennington, L. D., Aquila, B. M., Choi, Y., Valiulin, R. A. & Muegge, I. Positional Analogue 605 \nScanning: An Effective Strategy for Multiparameter Optimization in Drug Design. J Med 606 \nChem 63, 8956-8976, doi:10.1021/acs.jmedchem.9b02092 (2020). 607 \n22 Gaulton, A. et al. ChEMBL: a large-scale bioactivity database for drug discovery. Nucleic 608 \nAcids Res 40, D1100-1107, doi:10.1093/nar/gkr777 (2012). 609 \n23 Zdrazil, B.  et al.  The ChEMBL Database in 2023: a drug dis covery platform spanning 610 \nmultiple bioactivity data types and time periods. Nucleic Acids Res  52, D1180 -D1192, 611 \ndoi:10.1093/nar/gkad1004 (2024). 612 \n24 Muegge, I. & Hu, Y. In Silico Positional Analogue Scanning with Amber GPU -TI. J Chem 613 \nInf Model 62, 4448-4459, doi:10.1021/acs.jcim.2c00860 (2022). 614 \n25 Schönherr, H. & Cernak, T. Profound Methyl Effects in Drug Discovery and a Call for New 615 \nC\n  H Methylation Reactions. Angewandte Chemie International Edition 52, 12256-12267 616 \n(2013). 617 \n26 Cramer, J., Sager, C. P. & Ernst,  B. Hydroxyl groups in synthetic and natural -product-618 \nderived therapeutics: a perspective on a common functional group. J Med Chem 62, 8915-619 \n8930 (2019). 620 \n27 Chiodi, D. & Ishihara, Y. “Magic chloro”: profound effects of the chlorine atom in drug 621 \ndiscovery. J Med Chem 66, 5305-5331 (2023). 622 \n28 Gillis, E. P., Eastman, K. J., Hill, M. D., Donnelly, D. J. & Meanwell, N. A. Applications of 623 \nfluorine in medicinal chemistry. J Med Chem 58, 8315-8359 (2015). 624 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 30 \n29 Hopkins, A. L., Groom, C. R. & Alex, A. Ligand efficiency: a useful metric for lead selection. 625 \nDrug Discov Today 9, 430-431, doi:10.1016/S1359-6446(04)03069-7 (2004). 626 \n30 Zhu, T.  et al.  Hit identification and optimization in virtual screening: practical 627 \nrecommendations based on a critical literature analysis. J Med  Chem 56, 6560 -6572, 628 \ndoi:10.1021/jm301916b (2013). 629 \n31 Kuntz, I. D., Chen, K., Sharp, K. A. & Kollman, P. A. The maximal affinity of ligands. Proc 630 \nNatl Acad Sci U S A 96, 9997-10002, doi:10.1073/pnas.96.18.9997 (1999). 631 \n32 Halgren, T. A. Identifying and Char acterizing Binding Sites and Assessing Druggability. J 632 \nChem Inf Model 49, 377-389, doi:10.1021/ci800324m (2009). 633 \n33 Thomas, M., Bender, A. & de Graaf, C. Integrating structure -based approaches in 634 \ngenerative molecular design. Curr Opin Struct Biol  79, 10255 9, 635 \ndoi:10.1016/j.sbi.2023.102559 (2023). 636 \n34 Flowers, J. et al. Expanding automated multiconformer ligand modeling to macrocycles and 637 \nfragments. Elife 14, doi:10.7554/eLife.103797 (2025). 638 \n35 Di, L., Kerns, E. H. & Carter, G. T. Drug -like property concepts in pharmaceutical design. 639 \nCurr Pharm Des 15, 2184-2194, doi:10.2174/138161209788682479 (2009). 640 \n36 Morgan, P. et al. Can the flow of medicines be improved? Fundamental pharmacokinetic 641 \nand pharmacological principles toward improving Phase II survival. Drug Discov Today 17, 642 \n419-424, doi:10.1016/j.drudis.2011.12.020 (2012). 643 \n37 Houston, J. B. Utility of in vitro drug metabolism data in predicting in vivo metabolic 644 \nclearance. Biochem Pharmacol 47, 1469-1479, doi:10.1016/0006-2952(94)90520-7 (1994). 645 \n38 Iwatsubo, T. et al. Prediction of in vivo drug metabolism in the human liver from in vitro 646 \nmetabolism data. Pharmacol Ther  73, 147 -171, doi:10.1016/s0163 -7258(96)00184-2 647 \n(1997). 648 \n39 Amidon, G. L., Lennernas, H., Shah, V. P. & Crison, J. R. A theoretical basis for a 649 \nbiopharmaceutic drug classification: the correlation of in vitro drug product dissolution and 650 \nin vivo bioavailability. Pharm Res 12, 413-420, doi:10.1023/a:1016212804288 (1995). 651 \n40 Hopkins, A. L., Keseru, G. M., Leeson, P. D., Rees, D. C. & Reynolds, C. H. The role of 652 \nligand efficiency metrics in drug discovery. Nat Rev Drug Discov  13, 105 -121, 653 \ndoi:10.1038/nrd4163 (2014). 654 \n41 Young, R. J. & Leeson, P. D. Mapping the Efficiency and Physicochemical Trajectories of 655 \nSuccessful Optimizations. J Med Chem  61, 6421 -6467, 656 \ndoi:10.1021/acs.jmedchem.8b00180 (2018). 657 \n42 Rocklin, G. J., Mobley, D. L., Dill, K. A. & Hunenberger, P. H. Calculat ing the binding free 658 \nenergies of charged species based on explicit -solvent simulations employing lattice -sum 659 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 31 \nmethods: an accurate correction scheme for electrostatic finite -size effects. J Chem Phys 660 \n139, 184103, doi:10.1063/1.4826261 (2013). 661 \n43 Kim, H. et al. Artificial Intelligence in Drug Discovery: A Comprehensive Review of Data -662 \ndriven and Machine Learning Approaches. Biotechnol Bioprocess Eng  25, 895 -930, 663 \ndoi:10.1007/s12257-020-0049-y (2020). 664 \n44 Kumar, A., Kini, S. G. & Rathi, E. A Recent Appraisal of A rtificial Intelligence and In Silico 665 \nADMET Prediction in the Early Stages of Drug Discovery. Mini Rev Med Chem 21, 2788-666 \n2800, doi:10.2174/1389557521666210401091147 (2021). 667 \n45 Fu, L. et al. ADMETlab 3.0: an updated comprehensive online ADMET prediction platform 668 \nenhanced with broader coverage, improved performance, API functionality and decision 669 \nsupport. Nucleic Acids Res 52, W422-W431, doi:10.1093/nar/gkae236 (2024). 670 \n46 Myung, Y., de Sa, A. G. C. & Ascher, D. B. Deep -PK: deep learning for small molecule 671 \npharmacokinetic and toxicity prediction. Nucleic Acids Res  52, W469 -W475, 672 \ndoi:10.1093/nar/gkae254 (2024). 673 \n47 Swanson, K. et al. ADMET-AI: a machine learning ADMET platform for evaluation of large-674 \nscale chemical libraries. Bioinformatics 40, doi:10.1093/bioinformatics/btae416 (2024). 675 \n48 Bergstrom, C. A., Wassvik, C. M., Johansson, K. & Hubatsch, I. Poorly soluble marketed 676 \ndrugs display solvation limited solubility. J Med Chem  50, 5858 -5862, 677 \ndoi:10.1021/jm0706416 (2007). 678 \n49 Obach, R. S. Prediction of human clearance of twenty-nine drugs from hepatic microsomal 679 \nintrinsic clearance data: An examination of in vitro half -life approach and nonspecific 680 \nbinding to microsomes. Drug Metab Dispos 27, 1350-1359 (1999). 681 \n50 Friden, M. et al. Structure-brain exposure relationships in rat and human using a novel data 682 \nset of unbound drug concentrations in brain interstitial and cerebrospinal fluids. J Med 683 \nChem 52, 6233-6243, doi:10.1021/jm901036q (2009). 684 \n51 Fink, E. A.  et al.  Structure-based discovery of nonopioid analgesics actin g through the 685 \nalpha(2A)-adrenergic receptor. Science 377, eabn7065, doi:10.1126/science.abn7065 686 \n(2022). 687 \n52 Leung, C. S., Leung, S. S., Tirado-Rives, J. & Jorgensen, W. L. Methyl effects on protein -688 \nligand binding. J Med Chem 55, 4489-4500, doi:10.1021/jm3003697 (2012). 689 \n53 Sindt, F., Bret, G. & Rognan, D. On the Difficulty to Rescore Hits from Ultralarge Docking 690 \nScreens. J Chem Inf Model 65, 5553-5566, doi:10.1021/acs.jcim.5c00730 (2025). 691 \n54 Lyu, J. et al. Ultra-large library docking for discovering new chemotypes. Nature 566, 224-692 \n+, doi:10.1038/s41586-019-0917-9 (2019). 693 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 32 \n55 Moroz, Y., Chuprina, A. & Mykytenko, D. Enamine REAL DataBase - an instrumental and 694 \npractical vehicle for charting new regions of the relevant drug discovery chemical space . 695 \nAbstr Pap Am Chem S 251 (2016). 696 \n56 Verteramo, M. L. et al. Interplay of halogen bonding and solvation in protein-ligand binding. 697 \niScience 27, 109636, doi:10.1016/j.isci.2024.109636 (2024). 698 \n57 Krimmer, S. G., Betz, M., Heine, A. & Klebe, G. Methyl, ethyl, propyl, butyl: futile but not for 699 \nwater, as the correlation of structure and thermodynamic signature shows in a congeneric 700 \nseries of thermolysin inhibitors. ChemMedChem 9, 833-846, doi:10.1002/cmdc.201400013 701 \n(2014). 702 \n58 Wermuth, C. G. The Practice of Medic inal Chemistry.  (Elsevier Science & Technology, 703 \n2008). 704 \n59 Gahbauer, S. et al. Iterative computational design and crystallographic screening identifies 705 \npotent inhibitors targeting the Nsp3 macrodomain of SARS -CoV-2. Proceedings of the 706 \nNational Academy of Sciences 120, e2212931120 (2023). 707 \n60 Liu, F. et al. The impact of library size and scale of testing on virtual screening. Nature 708 \nChemical Biology, 1-7 (2025). 709 \n61 Clopper, C. J. & Pearson, E. S. The use of confidence or fiducial limits illustrated in the 710 \ncase of the binomial. Biometrika 26, 404-413 (1934). 711 \n62 Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat 712 \nMethods 17, 261-272 (2020). 713 \n63 Lu, C. et al. OPLS4: Improving Force Field Accuracy on Challenging Regimes of Chemical 714 \nSpace. J Chem Theory Comput 17, 4291-4300, doi:10.1021/acs.jctc.1c00302 (2021). 715 \n64 Balius, T. E.  et al.  Testing inhomogeneous solvation theory in structure -based ligand 716 \ndiscovery. Proceedings of the National Academy of Sciences 114, E6839-E6846 (2017). 717 \n65 Singh, I.  et al.  Structure-based discovery of conformationally selective inhibitors of the 718 \nserotonin transporter. Cell 186, 2160-2175. e2117 (2023). 719 \n66 Vigneron, S. F. et al. Docking 14 Million Virtual Isoquinuclidines against the μ and κ Opioid 720 \nReceptors Reveals Dual Antagonists–Inverse Agonists with Reduced Withdrawal Effects. 721 \nACS Central Science (2025). 722 \n67 Hua, T. et al. Activation and signaling mechanism revealed by cannabinoid receptor -Gi 723 \ncomplex structures. Cell 180, 655-665. e618 (2020). 724 \n68 Hua, T. et al. Crystal structures of agonist-bound human cannabinoid receptor CB1. Nature 725 \n547, 468-471 (2017). 726 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 33 \n69 Manglik, A. et al. Structure-based discovery of opioid analgesics with reduced side effects. 727 \nNature 537, 185-190 (2016). 728 \n70 Yung-Chi, C. & Prusoff, W. H. Relationship between the inhibition constant (KI) and the 729 \nconcentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction. 730 \nBiochemical pharmacology 22, 3099-3108 (1973). 731 \n71 Schuller, M. et al. Fragment binding to the Nsp3 macrodomain of SARS -CoV-2 identified 732 \nthrough crystallographic screening and computational docking. Science advances  7, 733 \neabf8711 (2021). 734 \n72 Maeda, S.  et al.  Development of an antibody fragment that stabilizes GPCR/G -protein 735 \ncomplexes. Nat Commun 9, 3712, doi:10.1038/s41467-018-06002-w (2018). 736 \n73 Goddard, T. D.  et al. UCSF ChimeraX: Meeting modern challenges in visualization and 737 \nanalysis. Protein Sci 27, 14-25, doi:10.1002/pro.3235 (2018). 738 \n74 Liebschner, D. et al. Macromolecular structure determination using X -rays, neutrons and 739 \nelectrons: recent developments in Phenix. Acta Crystallogr D Struct Biol  75, 861 -877, 740 \ndoi:10.1107/S2059798319011471 (2019). 741 \n75 Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coo t. 742 \nActa Crystallogr D Biol Crystallogr 66, 486-501, doi:10.1107/S0907444910007493 (2010). 743 \n76 Schuller, M. et al. Fragment binding to the Nsp3 macrodomain of SARS -CoV-2 identified 744 \nthrough crystallographic screening and computational docking. Sci Adv  7, 745 \ndoi:10.1126/sciadv.abf8711 (2021). 746 \n77 Correy, G. J.  et al.  Exploration of structure -activity relationships for the SARS -CoV-2 747 \nmacrodomain from shape -based fragment linking and active learning. Sci Adv  11, 748 \neads7187, doi:10.1126/sciadv.ads7187 (2025). 749 \n 750 \n  751 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 34 \nMethods  752 \nChEMBL database molecular pairs. From the PostgreSQL version of CHEMBL3423 753 \n(https://doi.org/10.6019/CHEMBL.database.34), we filtered the database for compounds with 754 \nactivities reported against a single protein, with either Ki, Kd, IC50 or EC50 as the activity type. 755 \nTo be considered a molecular pair, compounds had to have the same assay ID, the same 756 \ntarget ID, the same activity type, and the same reference document (publication) ID, in 757 \naddition to having a parent-analog relationship as defined in this work, meaning a C-H group 758 \nreplaced by C-OH, C-F, C-Cl, C-Br, C-CH3 or an aromatic carbon replaced by a nitrogen. This 759 \nleft us with 191,732 parent-analog pairs. 760 \n 761 \nAnalog enumeration and synthesis. Eighteen parent compounds were selected from 762 \npreviously published literature51,59,60 or datasets 763 \n(https://asapdiscovery.org/outputs/molecules/#ASAP-SARS-COV-2-NSP3-MAC1) based on 764 \ntheir known binding activity, structural relevance, or representation of diverse chemical 765 \nscaffolds. Starting from a parent compound, we used RDKit (www.rdkit.org) to identify all C-H 766 \nbonds and iteratively replace the hydrogen atom with a methyl, hydroxyl, fluoro, chloro or 767 \nbromo group. We also identified aromatic carbons with two heavy-atom neighbors and 768 \nreplaced them with nitrogen. Every analog generated was represented as a canonical 769 \nisomeric SMILES string and added to a set to remove duplicates. For each of the 18 parents, 770 \nthe full set of possible analogs that could be synthesized for < $400 for 10mg were ordered. 771 \n 772 \nConfidence intervals and statistics. When reported, 95% confidence intervals were 773 \nderived from bootstrap resampling with 10,000 iterations. In cases where the observed 774 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 35 \nfrequency of success is exactly 0, bootstrap will fail to give an upper bound for the interval. In 775 \nsuch cases, we used the Clopper-Pearson method to estimate an upper bound based on the 776 \nsample size.61 Pearson and Spearman correlations with their associated p-values were 777 \ncomputed using the scipy.stats module from the SciPy package62. 778 \n 779 \nFEP simulations. These were conducted using FEP+ within the Schrödinger software 780 \nsuite (versions 2025-2) with the OPLS4 force field63  and the modified SPC water model. The 781 \ndefault setting was used for the number of lambda windows selection where it depends on the 782 \ntype of perturbations; charge-changing, core hopping, and all other perturbations have 24, 16, 783 \nand 12 lambda windows respectively. For alchemical transformations with charge changes, 784 \nthe total charge of the simulation box was kept constant by transmuting a Na+ or Cl- ion to 785 \nwater or vice versa. In addition, a 0.15 M concentration of NaCl was added to the simulation 786 \nbox of charge-changing perturbations. For α2A, CB2, and SERT, the FEP+ membrane 787 \nprotocol was applied where a POPC membrane was added to the system in simulation. All 788 \nother settings were kept default except the simulation time was extended from 5 ns to 10 ns.  789 \n 790 \nThe default FEP map generation protocol was used with the parent compound selected 791 \nas the biased node. In preparing proteins and ligands for FEP+, the Schrödinger protein 792 \npreparation workflow and LigPrep were used. The initial binding poses of parent compo unds 793 \nwere from poses generated by DOCK3.8, and analogs were aligned to the parents with severe 794 \nsteric clashes resolved using the FEP+ Pose Builder workflow. 795 \n 796 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 36 \nDocking: Whereas no structural information was used in the design of analogs from 797 \nparent compounds, for FEP+ calculations and for post hoc structural analysis, we generated 798 \nligand bound complexes of the parents and relevant ligands using DOCK.3.854,64. Ligands were 799 \ndocked into the receptor binding site using grids prepared in previous studies51,59,60,65,66. 800 \n 801 \n Assay selection. We use one consistent readout per parent series (no mixing within a 802 \nseries). SERT and AmpC are reported as Ki; Mac1 is reported as IC50 from peptide 803 \ndisplacement (a scalable functional hydrolysis assay is unavailable); GPCR agonist series are 804 \nreported as EC50 and antagonist series as Ki, such that MOR is EC50 -only whereas α2A and 805 \nCB2 include EC50 and Ki depending on the parent series.  Representative concentration –806 \nresponse curves and the corresponding Z values for each assay are provided in 807 \nSupplementary Figure 3. 808 \n 809 \nTransporter assays for SERT Ki.  SERT activity was measured using the 810 \nNeurotransmitter Transporter Uptake Assay Kit from Molecular Devices (Catalog #R8174), 811 \nfollowing the manufacturer’s protocol with slight modifications as described previously65. 812 \nHEK293 cells stably expressing human SERT were plated in poly-L-lysine (PLL)-coated 384-813 \nwell black, clear-bottom plates at a density of 15,000 cells in 40 µL per well, using DMEM 814 \nsupplemented with 1% dialyzed FBS (dFBS). Cells were incubated overnight at 37°C with 5% 815 \nCO₂ to allow adherence and recovery. The following day, the medium was carefully aspirated, 816 \nand cells were incubated with 25 µL per well of test compound solutions prepared in assay 817 \nbuffer (1× HBSS, 20 mM HEPES, pH 7.4, supplemented with 1 mg/mL BSA) for 30 minutes at 818 \n37°C. After drug treatment, 25 µL per well of dye solution (as provided in the kit) was added 819 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 37 \ndirectly to the wells, followed by an additional 30-minute incubation at 37°C. Fluoxetine (10 820 \nµM) was used as a positive control for SERT inhibition. Fluorescence was measured using 821 \nthe FlexStation II microplate reader with excitation at 440 nm and emission at 520 nm. 822 \nRelative fluorescence units (RFUs) were exported and analyzed using GraphPad Prism 10.0 823 \nto calculate IC₅₀ values, from which Ki values were derived using the Cheng-Prusoff equation. 824 \n 825 \nCB2 radioligand binding assay. CB2 receptor binding assays were performed using 826 \nmembrane preparations from HEK293 cells stably expressing human CB2, following 827 \npreviously published methods67,68. Membranes were resuspended in TME buffer containing 828 \n0.1% BSA (w/v) and 25 µg of membrane protein was added per well. The assay was 829 \nconducted using the radioligand [³H]CP-55,940 at a final concentration of 0.75 nM, prepared 830 \nin assay buffer. Nonspecific binding was defined in the presence of 5 µM unlabeled CP-831 \n55,940. Test compounds were applied at increasing concentrations to assess competition. 832 \nReactions were incubated at 30 °C for 1 hour with gentle shaking. After incubation, samples 833 \nwere transferred to Unifilter GF/B 96-well filter plates and filtered using a Packard Filtermate-834 \n196 cell harvester (PerkinElmer). Plates were washed four times with ice-cold wash buffer (50 835 \nmM Tris-HCl, 5 mM MgCl₂, 0.5% BSA, pH 7.4). Radioactivity bound to the filters was 836 \nquantified via liquid scintillation counting. Specific binding was calculated by subtracting 837 \nnonspecific from total binding. IC₅₀ and Ki values were calculated using nonlinear regression 838 \nin GraphPad Prism 9 using the Cheng-Prusoff equation. 839 \n 840 \nα2A Receptor Binding Assay.  α2A adrenergic receptor binding was performed using 841 \nmembrane preparations from insect cells expressing human α2A receptors, as previously 842 \ndescribed51. Membranes were incubated with increasing concentrations of test compounds 843 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 38 \nand 5 nM [³H]-rauwolscine in buffer containing 20 mM HEPES (pH 7.5) and 100 mM NaCl, at 844 \nroom temperature for 2 hours. After incubation, samples were filtered onto GF/B filter plates, 845 \nwashed with ice-cold buffer, and radioactivity was quantified by liquid scintillation counting.  846 \nIC₅₀ and Ki values were derived using nonlinear regression in GraphPad Prism. 847 \n 848 \n 849 \nGloSensor cAMP assay for α2A, CB2, and MOR.  The GloSensor cAMP assay was 850 \nperformed following the manufacturer’s instructions (Promega) with slight modifications as 851 \npreviously described51,69. Briefly, wild-type (WT) human α2A, CB2, and MOR were cloned into 852 \nthe pcDNA3.1 vector and co-transfected with the 22F cAMP GloSensor plasmid into 853 \nHEK293T cells cultured in 6-well plates. After 24 hours, cells were reseeded into 96-well white 854 \nplates in CO₂-independent medium and equilibrated with GloSensor cAMP reagent as per the 855 \nmanufacturer's protocol. Cells were incubated for 1 hour at 37 °C followed by 1 hour at room 856 \ntemperature. Where applicable, 10 μM forskolin was used to elevate basal cAMP levels for 857 \nassessing receptor-mediated inhibition. Serially diluted test compounds were added, and 858 \nluminescence signals were recorded using a PerkinElmer microplate reader. Data were 859 \nanalyzed using GraphPad Prism 9.0 to calculate EC₅₀ or IC₅₀ values. 860 \n 861 \nAmpC β-lactamase Inhibition Assay. The AmpC β-lactamase inhibition assay was 862 \nperformed as previously described60. The candidate inhibitors were dissolved in DMSO (20 mM stock) 863 \nand diluted to maintain a constant 1% DMSO (v/v) in 50 mM sodium cacodylate buffer (pH 6.5). 864 \nAssays were performed in the presence of 0.01% Triton X-100 to reduce aggregation artifacts. AmpC 865 \nenzymatic activity was monitored spectrophotometrically using CENTA or nitrocefin as substrates.  866 \nInitial screening was performed at 200 µM, 100 µM, and 40 µM compound concentrations. Substrate 867 \nconcentrations were selected based on known Km values to achieve defined [S]/Km ratios: for CENTA 868 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 39 \n([S] = 50 µM, Km = 27.6 µM) and nitrocefin ([S] = 100 µM or 28 µM, Km = 180 µM). Reactions were 869 \ncarried out in 96-well format on a BMG Labtech CLARIOstar plate reader, with substrate and enzyme 870 \ninjected into wells containing the inhibitor, followed by kinetic measurement over 50 seconds.  IC₅₀ 871 \nvalues were determined by fitting inhibition curves in GraphPad Prism using a fixed Hill coefficient of 1, 872 \nand Ki values were calculated using the Cheng-Prusoff equation70.  873 \n 874 \n 875 \nHTRF assay for Mac1.  Binding of the compounds to Mac1 was assessed by the 876 \ndisplacement of an ADPr conjugated biotin peptide from His6-tagged protein using a HTRF- 877 \nbased assay, as previously described71.The expression sequences used for SARS-CoV-2 878 \nMac1 are listed below. All proteins were expressed and purified as described previously for 879 \nSARS-CoV-2 Mac171. Compounds were dispensed into ProxiPlate-384 Plus (PerkinElmer) 880 \nassay plates using an Echo 650 Liquid Handler (Beckman Coulter). Binding assays were 881 \nconducted in a final volume of 16 μl with 12.5 nM NSP3 Mac1 protein, 200 nM peptide 882 \nARTK(Bio)QTARK(Aoa- RADP)S (Cambridge Peptides), 1:20000 Anti-His6-Eu3+ cryptate 883 \n(HTRF donor, PerkinElmer AD0402) and 1:500 Streptavidin-XL665 (HTRF acceptor, 884 \nPerkinElmer 610SAXLB) in assay buffer (25 mM 4-(2-hydroxyethyl)-1-piperazi‐ 885 \nneethanesulfonic acid (HEPES) pH 7.0, 20 mM NaCl, 0.05% bovine serum albumin and 886 \n0.05% Tween-20, the latter also to reduce aggregation artifacts). Assay reagents were 887 \ndispensed manually into plates using an electronic multichannel pipette. Mac1 and peptide 888 \nwere preincubated for 30 min at room temperature before HTRF reagents were added. 889 \nFluorescence was measured after a 1 hour incubation at room temperature using a Perkin 890 \nElmer EnVision 2105-0010 Dual Detector Multimode microplate reader with dual emission 891 \nprotocol (A = excitation of 320 nm, emission of 665 nm, and B = excitation of 320 nm, 892 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 40 \nemission of 620 nm). Compounds were tested in triplicate in a 14-point dose response. Raw 893 \ndata were processed to give an HTRF ratio (channel A/B × 10,000), which was used to 894 \ngenerate IC50 curves using nonlinear regression using GraphPad Prism v.10.0.2 (GraphPad 895 \nSoftware, CA, USA). 896 \n 897 \nhERG channel inhibition. hERG channel inhibition was evaluated using a Thallium 898 \nFlux assay on HEK293 cells stably expressing the human Ether-à-go-go Related Gene 899 \n(hERG) potassium channel. Cells were seeded at a density of 8000 cells per well in 384-well 900 \npoly-D-lysine–coated plates and incubated for 24 hours under standard conditions (37 °C, 5% 901 \nCO₂). The next day, a thallium-sensitive dye was added to the cells, followed by a 1-hour 902 \nincubation to ensure dye uptake. Test compounds were added to achieve a final 903 \nconcentration of 30 μM in 0.5% DMSO, and the cells were incubated for an additional 30 904 \nminutes at room temperature. Subsequently, a stimulation buffer containing thallium was 905 \nadded and fluorescence measurements were taken using a FLIPR Tetra system. Data were 906 \ncollected every 3 seconds for 3 minutes (excitation: 470–495 nm, emission: 515–575 nm). 907 \nFluorescence intensity over time was analyzed to calculate the area under the curve (AUC)  908 \nfrom which percentage inhibition was determined versus haloperidol at 100μM (positive 909 \ncontrol) and vehicle (DMSO). 910 \n 911 \nα2A Receptor Purification and Structure Determination.  Wild-type human α2A 912 \nadrenergic receptor (α2AAR) was cloned into a pVL1392 vector with an N -terminal FLAG tag. 913 \nThe construct was expressed in Spodoptera frugiperda (Sf9) insect cells using the BestBac 914 \nsystem. Cells at a density of 4 × 10⁶ cells/mL were infected with virus and incubated for 48 hours 915 \nat 27 °C. The receptor was solubilized and purified by FLAG affinity chromatography and size-916 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 41 \nexclusion chromatography in the presence of 10 μM compound ‘4905. Monomeric peak 917 \nfractions were concentrated used for G protein complex formation.  GαoGβ1γ2 heterotrimeric G 918 \nproteins were expressed in Hi5 insect cells and purified using Ni² ⁺-affinity following detergent 919 \nsolubilization and dephosphorylation. The final α 2AAR–GαoGβ1γ2–scFv1672 complex was 920 \nassembled in the presence of ‘4905 and purified by size -exclusion chromatography. Cryo-EM 921 \ngrids were prepared using UltrAufoil R1.2/1.3 300 mesh grids and vitrified in liquid ethane. Data 922 \nwere collected on a Titan Krios G3 electron microscope equipped with a K3 direct electron 923 \ndetector. Image processing was performed using cryoSPARC, yielding a final reconstruction at 924 \n~2.8 Å resolution. Model building and refinement was performed using PDB 7EJ8 as a starting 925 \nmodel using ChimeraX73, Phenix74, and Coot75. The final structures have been deposited in the 926 \nPDB with accession codes 9PLO and 9PLN.  927 \n 928 \nMac1 Purification and Crystallization.  Wild-type Mac1 protein (P43 construct, 929 \nresidues 3–169) was expressed in E. coli BL21(DE3) as an N-terminal His₆-tagged construct 930 \nand purified by Ni²⁺ affinity chromatography 76. The His-tag was cleaved with TEV protease, 931 \nfollowed by size-exclusion chromatography (Superdex 75) in 20 mM Tris-HCl (pH 7.5), 932 \n150 mM NaCl, and 1 mM DTT. Purified protein was concentrated to 40 mg/ml for 933 \ncrystallization and stored at -80°C. 934 \n 935 \nCrystals were obtained by sitting-drop vapor diffusion in 28% PEG 3000 and 100 mM 936 \nCHES (pH 9.5). Compounds (100 mM in DMSO) were added to crystal drops using an Echo 937 \n650 acoustic dispenser to a final concentration of 10 mM. Crystals were incubated at room 938 \ntemperature for 2–4 hours and vitrified in liquid nitrogen without additional cryoprotection. X-939 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 42 \nray diffraction data were collected at ALS (beamline 8.3.1) and processed using XDS and 940 \nAimless. Structures were determined to resolutions ranging from 0.97 to 1.02 Å. Ligands with 941 \nlow occupancy or conformational disorder were modeled using PanDDA and refined using 942 \nPhenix 77.PanDDA event maps are shown in Supplementary Figure 7. The structure of ‘3184 943 \nwas determined using a racemic preparation of the compound (‘9037). X-ray data collection 944 \nand refinement statistics are summarized in SI Table 2, as are the 32 PDB IDs.   945 \n 946 \nMicrosomal Stability.  Microsomal stability of compounds was evaluated using pooled 947 \nmouse liver microsomes (XenoTech, M3000/lot #2010026) to estimate their metabolic stability 948 \nand predict hepatic clearance. Each compound was incubated at 2 μM in a reaction mixture 949 \ncontaining 0.42 mg/mL microsomal protein, phosphate buffer (100 mM, pH 7.4), MgCl₂ (3.3 950 \nmM), NADPH (3 mM), glucose-6-phosphate (5.3 mM), and glucose-6-phosphate 951 \ndehydrogenase (0.67 units/mL). Reactions were conducted at 37 °C in 96-well plates with 952 \nshaking at 100 rpm. Samples were collected at five time points (0, 7, 15, 25, and 40 minutes), 953 \nand reactions were quenched by adding five volumes of acetonitrile containing an internal 954 \nstandard. After centrifugation at 5500 rpm for 5 minutes, the supernatants were analyzed via 955 \nHPLC-MS/MS. The elimination rate constant (kel), half-life (t₁/₂), and intrinsic clearance (Clint) 956 \nwere calculated by plotting the natural logarithm of the remaining parent compound versus 957 \ntime. Stability was compared with reference standards such as imipramine and propranolol. 958 \n 959 \nPlasma Protein Binding (PPB). Plasma protein binding (PPB) was measured using 960 \nequilibrium dialysis with a 14 kDa molecular weight cut-off membrane in a 96-well HTD96b 961 \ndialyzer. Mouse plasma containing 1 μM test compound (0.005% DMSO, 1% acetonitrile) was 962 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 43 \nplaced in one chamber, and phosphate-buffered saline (PBS, pH 7.4) in the opposing 963 \nchamber. The assembled plates were incubated at 37 °C with 5% CO₂ and ∼95% humidity, 964 \nshaking at 250 rpm for 5 hours to reach equilibrium. After incubation, aliquots from each 965 \nchamber were mixed with equal volumes of the blank opposite matrix and processed with 966 \nacetonitrile containing internal standard. Supernatants obtained after centrifugation were 967 \nanalyzed by HPLC-MS/MS. The percentage of compound bound to plasma proteins was 968 \ncalculated using the peak area ratio in buffer to plasma compartments. Recovery and stability 969 \nstandards were included to ensure accuracy and reliability. Verapamil served as a reference 970 \ncontrol. Most compounds showed moderate binding, typically ranging between 75–85%. 971 \n 972 \nPlasma stability. Plasma stability was assessed in non-sterile mouse plasma (Li-973 \nheparin treated) at 1 μM concentration (final DMSO content was 0.005%). Incubations were 974 \ncarried out in aliquots of 60 μL each (two per time point) at 37 °C under 5% CO₂ and high 975 \nhumidity (∼95%). The reactions were quenched with 240 µL of 90% acetonitrile containing an 976 \ninternal standard at 0, 20, 40, 60, and 120 minutes, followed by centrifugation at 6000 rpm for 977 \n5 minutes. Supernatants were analyzed via HPLC-MS/MS to determine the percentage of 978 \nparent compound remaining at each time point. Data were plotted to calculate compound half-979 \nlives (t₁/₂). Reference compounds, verapamil and propantheline, were used as high and low 980 \nstability controls, respectively. This assay is critical for identifying compounds susceptible to 981 \ndegradation by plasma esterases or hydrolytic enzymes, helping inform pharmacokinetic 982 \noptimization strategies during lead selection. 983 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 44 \n 984 \nThermodynamic Solubility. Aqueous thermodynamic solubility was determined in 985 \nPBS (pH 7.4) for 247 compounds using a shake-flask method followed by UV absorbance 986 \nquantification. Dry powder compounds were dissolved in PBS to a theoretical concentration of 987 \n4 mM and incubated in duplicates at 25 °C for 4 and 24 hours with shaking. After incubation, 988 \nsamples were filtered using HTS 96-well filter plates. The filtrates were diluted 2-fold in 989 \nacetonitrile with 4% DMSO for UV analysis. The incubation samples for charged molecules 990 \nwere additionally diluted 10-fold with 50% acetonitrile/PBS with 2% final DMSO. Calibration 991 \ncurves (0–200 μM) were prepared in 50% acetonitrile/PBS (2% final DMSO). Absorbance was 992 \nmeasured between 230–550 nm using a SpectraMax Plus microplate reader. Compound-993 \nspecific absorbance maxima were used to calculate concentration using SoftMax Pro and 994 \nExcel. The assay dynamic range is ~2–400 μM (~20–4000 μM for charged molecules), with 995 \nvalues near the upper limit treated as semi-quantitative. Ondansetron was used as a 996 \nreference compound. This method reflects equilibrium solubility under physiologically relevant 997 \nconditions and helps rank compounds for formulation feasibility. 998 \n 999 \nPAMPA-BBB. Passive blood-brain barrier (BBB) permeability was estimated using a 1000 \nParallel Artificial Membrane Permeability Assay (PAMPA-BBB) with a phospholipid-coated 1001 \nmembrane simulating the brain endothelium. Test compounds (50 μM in Prisma HT buffer, pH 1002 \n7.4, with 0.5% DMSO) were added to donor wells, while brain sink buffer was added to the 1003 \nacceptor wells. The donor and acceptor chambers were separated by a 0.45 µm filter 1004 \nmembrane coated with brain polar lipids. Plates were incubated without agitation at room 1005 \ntemperature for 4 hours. Post-incubation, samples from both chambers, as well as a standard 1006 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 45 \nsolution, were diluted with acetonitrile containing an internal standard. Apparent permeability 1007 \ncoefficients (log Papp) were calculated based on peak area ratio. Clozapine and 1008 \nchlorpromazine were used as high-permeability controls, while ranitidine represented low 1009 \npermeability. The assay provides a high-throughput, non-cell-based method for estimating 1010 \nCNS exposure potential. 1011 \n 1012 \nBehavioral analyses. All animal behavior experiments were conducted with the 1013 \nexperimenter blinded to treatment. Mice were habituated individually in Plexiglas enclosures 1014 \nfor 1 hour prior to testing. Compounds were administered subcutaneously 30 minutes before 1015 \nbehavioral assessment, and where applicable the α2A adrenergic receptor antagonist 1016 \natipamezole (2 mg/kg, intraperitoneally) was given 15 minutes before compound injection. Tail 1017 \nflick latency was measured by immersing the distal third of the tail in a 50°C water bath and 1018 \nrecording the withdrawal time. For the neuropathic pain model, spared nerve injury (SNI) was 1019 \nperformed under isoflurane anesthesia by ligating and transecting two of the three branches 1020 \nof the sciatic nerve, sparing the sural nerve. Mechanical thresholds were assessed 7 to 14 1021 \ndays post-surgery using von Frey filaments and the up-down method, and values were 1022 \nnormalized to each animal’s baseline. Thermal nociception was evaluated using a 55°C hot 1023 \nplate, and the latency to nocifensive behavior (paw lick or jump) was recorded with a cutoff of 1024 \n45 seconds to prevent tissue damage. Behavioral data are summarized as group size (n), 1025 \nmean, and SD (Supplementary Table 12). For the SNI experiment, we used two-way 1026 \nANOVA followed by Tukey’s post-hoc multiple comparisons. For the hotplate assay, we used 1027 \none-way repeated-measures ANOVA with Friedman post-hoc testing. For tail-flick, the 4905 1028 \ndose–response was analyzed by two-way ANOVA with Dunnett’s post-hoc comparisons to 1029 \nvehicle/control, whereas PS75 and 3629 dose groups were analyzed by one-way ANOVA 1030 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 46 \nwith Kruskal–Wallis post-hoc testing (Supplementary Table 12).For sample size，we did not 1031 \nperform formal a priori power calculations. Instead, group sizes were guided by our prior 1032 \nexperience with these assays and by precedent in the literature. This approach may limit 1033 \nsensitivity to small effects. Raw data for the animal assays are provided in Supplementary 1034 \nTable 13. 1035 \n 1036 \nData availability.  Most primary data is available through this manuscript and its extended 1037 \nand supplementary materials.  Crystal and cryoEM structures and supporting electron density 1038 \nare being made available via the Protein Data Bank, including PDB 9PLN, the structure of the 1039 \n2a/’4905 complex and PDB 14AB, 7IIW, 7IIX, 7IIY, 7IIZ, 7IJ0, 7IJ1, 7IJ2, 7IJ3, 7IJ4, 7IJ5, 7IJ6, 1040 \n7IJ7, 7IJ8, 7IJ9, 7IJA, 7IJB, 7IJC, 7IJD, 7IJE, 7IJF, 7IJG, 7IJH, 7IJI, 7IJJ, 7IJK, 7IJL, 14AM, 14AN, 1041 \n14AO, 14AP, and 7IJM, structures of Mac1 in complex with inhibitors.. 1042 \n 1043 \nCode availability. ZINC tools used in the selection of the small perturbation analogs are 1044 \nopenly available and zinc22.docking.org.   1045 \n 1046 \nAcknowledgements: Supported by US DARPA grant HR0011-19-2-0020 (to B .K.S., A .M., 1047 \nA.I.B., and B.L.R.), by US NIH R35GM122481 (to B.K.S.) by US ARPA-H grant 1AY1AX000035 1048 \n(PI J.S.F.), O.M. partially supported by US NIH postdoctoral fellowship F32GM154469.  We 1049 \nthank Yeyue Xiong for help with the FEP studies.  We thank the UCSF Cryo-EM facility staff for 1050 \ntraining and technical assistance. UCSF Cryo-EM equipment is partially supported by NIH 1051 \ngrants S10OD020054, S10OD021741 , and S10OD026881 , and by the Howard Hughes 1052 \nMedical Institute. 1053 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 47 \n 1054 \nAuthor contributions: B.K.S., X.X., and O.M. designed the project. X.X. and O.M. designed 1055 \nthe analogs with help from N.L. Alpha₂ receptor studies were performed by X.X. and H.H., with 1056 \nguidance from P.G. Alpha2A -related behavioral analyses were conducted by J.B., with 1057 \nguidance from A.I.B. Alpha₂ structural studies were performed by K.S., under the supervision 1058 \nof A.M. SERT binding studies were carried out by X.P.H. and J.W., with guidance from B.L.R. 1059 \nand ligand choice from Y.W. Mac1 biochemical assays were conducted by Y.U.D. M.S and M.D. 1060 \nsupervised by A.A., and Mac1 crystallography was performed by G.C. with guidance from J.S.F. 1061 \nCB2 receptor assays were performed by X.X. and C.T., with ligand advice from M.R. AmpC 1062 \nassays were conducted by X.X. and F.L. All in vitro ADME and safety assay s were supported 1063 \nby Y.H. and Y.K. through Bienta. FEP calculations and analysis were performed by DS guided 1064 \nby GZ and RA . O.M. performed statistical analyses for both ChEMBL  and experimental data. 1065 \nX.X., O.M., and B.K.S. prepared the manuscript. B.K.S. supervised the project. All authors 1066 \nreviewed and approved the final manuscript. 1067 \n 1068 \nCompeting interests B.K.S. is co-founder of Epiodyne, BlueDolphin, and Deep Apple 1069 \nTherapeutics, and serves on SAB for Schrodinger LLC, Vilya Therapeutics, Frontier Discovery 1070 \nLtd, and on the SRB of Genentech . B.L.R. is founder of Onsero Therapeutics. J.S.F. is a 1071 \nconsultant to and a shareholder of Vilya Therapeutics and Relay Therapeutics. The remaining 1072 \nauthors declare no competing interest.  1073 \n 1074 \n  1075 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 48 \nExtended Data Fig. 1 1076 \n 1077 \nExtended Data Fig. 1 Frequency of activity changes from single-atom substitutions in 1078 \nChEMBL and in this study.  a, Cumulative frequency of single non-hydrogen atom 1079 \nsubstitutions (CH3, F, Cl, Br, N, OH) in ChEMBL that improve activity, stratified by that of the 1080 \nparent compound. “Potent” denotes parents with activity ≤32 nM, “mid” 32 nM–1 μM, and 1081 \n“weak” 1 μM–1 mM.  b, For each target in this study, the percentage of analogs that improve 1082 \nor decrease activity by ≥3-fold or ≥10-fold versus their parent. Fold changes were rounded to 1083 \nthe nearest integer prior to thresholding (e.g., 9.6–9.9 counted as 10-fold). Error bars denote 1084 \n95% bootstrap confidence intervals (percentile method; 20,000 resamples). Compounds 1085 \nyielding no measurable curves were counted as ≥10-fold decreases. c, Parent-level summary 1086 \nof analog effects in this study (rounded fold-change).  1087 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 49 \nExtended Data Fig. 2 1088 \n  1089 \n 1090 \nExtended Data Fig. 2 Overview of Parent Compounds' Structures and Properties. 1091 \na. 2D structures of the parent compounds.  b. Molecular Weight (MW) and Affinity/potency 1092 \nvalues for the parent compounds. c. Structural Similarity matrix showing the relationships 1093 \nbetween parent compounds based on molecular fingerprint comparisons.   1094 \n 1095 \n  1096 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 50 \nExtended Data Fig. 31097 \n 1098 \nExtended Data Fig. 3 Correlation between parent compound or binding pocket 1099 \nproperties and affinity/potency fold change. 1100 \na. Correlation between parent molecular weight (MW) and log₁₀ affinity/potency fold 1101 \nchange 1102 \nb. Correlation between parent pKi and log₁₀ affinity/potency fold change.  1103 \nc. Correlation between binding pocket volume and log₁₀ affinity/potency fold change.  1104 \nd. Correlation between SiteMap Dscore and log₁₀ affinity/potency fold change.  1105 \n  1106 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 51 \nExtended Data Fig. 4  1107 \n 1108 \nExtended Data  Fig. 4 Correlations between physicochemical descriptors and 1109 \nΔpKi/pEC50 (log M) .  a. Scatter plot of ΔpKi  or pEC50 (log M) versus cLogP (RDKit). Each 1110 \npoint represents one analog; the solid line indicates a linear fit. Pearson’s R and Spearman’s ρ 1111 \nare shown. b. ΔpKi or pEC50 (analog − parent) against ΔcLogP for all analog–parent pairs with 1112 \n≥3× rounded potency/affinity improvement; ≥10× improvements are highlighted in red. c. 1113 \nScatter plot of ΔpKi or pEC50 (log M) versus ΔLE (analog − parent), where LE = 1.37 × pKi or 1114 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 52 \npEC50 / Nheavy. Pearson’s R and Spearman’s ρ are shown.  d. Scatter plot of ΔpKi or pEC50 1115 \n(log M) versus ΔLipE (analog − parent), where LipE = pKi or pEC50 − cLogP. Pearson’s R and 1116 \nSpearman’s ρ are shown.  e. Binding free -energy change versus heavy -atom change for 1117 \nimproved analogs. The y -axis shows ΔΔG (kcal/mol), computed from the affinity metric 1118 \n(Ki/IC50/EC50) as ΔG = RT ln(K) (298 K) and ΔΔG = ΔG_analog − ΔG_parent; negative ΔΔG 1119 \nindicates improved binding. The x -axis is ΔN (a nalog − parent; heavy -atom count difference). 1120 \nPoints with rounded fold ≥10 are highlighted in red; rounded 3 –9 are shown in blue.  f. 1121 \nAffinity/potency gains versus changes in lipophilic efficiency. Scatter plot of ΔpKi or pEC50 1122 \n(analog − parent) against ΔLipE (analog − parent) for all analog–parent pairs with ≥3× rounded 1123 \npotency/affinity improvement; ≥10× improvements are highlighted in red. 1124 \n 1125 \n  1126 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 53 \nExtended Data Fig. 5 1127 \n 1128 \nExtended Data Fig. 5 Structural variability caused by single-atom modifications 1129 \ncomplicates prediction of affinity/potency changes.  a. Overlay of the X-ray crystal 1130 \nstructures for 12 ligand-bound complexes from the ‘3453 series showing that single-atom 1131 \nmodifications can result in substantial ligand movement or induce multiple binding poses, 1132 \ncomplicating structure-based analysis and predictions.  b. Example of a binding pose shift: 1133 \ncompounds ‘9249, ‘3176 and ‘3194 adopt markedly different pose compared to the ‘3454 1134 \nparent. Two conformations were identified for ‘3194.  1135 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 54 \nExtended Data Fig. 6 1136 \n 1137 \nExtended Data Fig. 6 Single-atom substitutions frequently worsen PK properties a.Fraction of 1138 \nsingle-atom substitutions with ≥3-fold decreases in PK properties. Error bars represent 95% confidence 1139 \nintervals, estimated from 10,000 iterations of bootstrap resampling.b. Counts of analogs with measured 1140 \nPK parameters and >3-fold or >10-fold losses. 1141 \n  1142 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 55 \n 1143 \nExtended Data Fig. 7 1144 \n 1145 \nExtended Data Fig. 7 Pharmacokinetic profiles of 29 individual analogs with >10-fold 1146 \naffinity/potency improvement.  Fold-changes in in vitro pharmacokinetic properties, 1147 \nincluding microsomal stability, plasma fraction unbound, PAMPA permeability, plasma 1148 \nstability, solubility, and hERG inhibition, for 29 analogs showing >10-fold affinity/potency 1149 \nenhancements.  1150 \n 1151 \n 1152 \n 1153 \n1154 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 56 \nExtended Data Fig. 8 1155 \n 1156 \nExtended Data  Fig. 8 Distribution and cumulative effects of atom modifications on 1157 \nmolecular properties.  a. Density plots showing the distribution of log ₁₀ fold improvement for 1158 \ndifferent properties after one -atom modifications (Chlorine, Fluorine, Hydroxyl, Methyl, and 1159 \nNitrogen). Each row represents a property (e.g., Affinity/potency, Microsome stability, Solubility), 1160 \nand each column corresponds to a specifi c modification. Vertical dashed lines indicate 0 (no 1161 \nchange) and ±0.5 log ₁₀ fold change (approximately 3 -fold improvement or reduction).  b. 1162 \nCumulative distribution plots of the same data as in (a), showing the proportion of analogs 1163 \nachieving various levels of improvement or reduction. This visualization allows for comparison 1164 \nof the shift in property distributions across different modifications. 1165 \n 1166 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\nExtended Data Table1 FEP+ predictions and experimental activity readouts for analogs. 1167 \nName modification protein affinity/potency- \nfold \nAffinity/potenc\ny_uM \nexp_G Parent exp \nG \nparent \nname \nexp_G FEP_G fep_unc fep_G fep_G \nerr_abs \nfep_G \nerr_abs \nZ8771520648 C MOR 13.39 0.00 -12.50 -10.96 Z4407498716 -1.54 -11.85 0.41 -0.89 0.65 0.65 \nZ8771520652 C MOR 0.12 0.08 -9.69 -10.96 Z4407498716 1.27 -8.95 0.59 2.01 0.74 0.74 \nZ8771520655 C MOR 0.53 0.02 -10.59 -10.96 Z4407498716 0.37 -10.98 0.40 -0.02 0.39 0.39 \nZ8771520657 C MOR 0.13 0.07 -9.76 -10.96 Z4407498716 1.20 -11.81 0.41 -0.85 2.05 2.04 \nZ8771520658 C MOR 0.48 0.02 -10.53 -10.96 Z4407498716 0.44 -11.14 0.41 -0.18 0.61 0.61 \nZ8771520665 O MOR 0.07 0.14 -9.36 -10.96 Z4407498716 1.60 -12.00 0.50 -1.04 2.64 2.64 \nZ8771520670 O MOR 1.24 0.01 -11.09 -10.96 Z4407498716 -0.13 -11.19 0.50 -0.23 0.10 0.10 \nZ8771520680 F MOR 0.73 0.01 -10.77 -10.96 Z4407498716 0.19 -9.96 0.40 1.00 0.81 0.81 \nZ8771520687 F MOR 0.67 0.01 -10.72 -10.96 Z4407498716 0.24 -10.77 0.41 0.19 0.05 0.05 \nZ8771520688 F MOR 0.26 0.04 -10.17 -10.96 Z4407498716 0.80 -11.38 0.41 -0.42 1.21 1.21 \nZ8774357908 O MOR 0.77 0.01 -10.81 -10.96 Z4407498716 0.16 -9.40 0.45 1.56 1.41 1.41 \nZ8793812850 C MOR 0.93 0.01 -10.92 -10.96 Z4407498716 0.04 -8.62 0.75 2.34 2.30 2.30 \nZ8793812851 C MOR 0.82 0.01 -10.84 -10.96 Z4407498716 0.12 -10.85 0.54 0.11 0.01 0.00 \nZ8798881729 C MOR 10.80 0.00 -12.37 -10.96 Z4407498716 -1.41 -11.39 0.75 -0.43 0.98 0.98 \nZ8812737822 Cl MOR 0.04 0.26 -8.98 -10.96 Z4407498716 1.98 -10.86 0.41 0.10 1.88 1.88 \nZ8812742460 Cl MOR 1.67 0.01 -11.27 -10.96 Z4407498716 -0.30 -11.89 0.54 -0.93 0.62 0.62 \nZ1800559918 Cl alpha2 0.00 6.84 -7.05 -12.09 Z2750653629 5.04 -13.03 1.14 -0.94 5.98 5.98 \nZ2750652484 N alpha2 0.01 0.16 -9.28 -12.09 Z2750653629 2.81 -8.88 1.52 3.21 0.40 0.40 \nZ4767467064 F alpha2 18.39 0.00 -13.81 -12.09 Z2750653629 -1.73 -12.37 1.28 -0.28 1.44 1.44 \nZ4767467067 N alpha2 0.01 0.09 -9.59 -12.09 Z2750653629 2.50 -11.33 1.98 0.76 1.74 1.73 \nZ4767467070 N alpha2 0.01 0.14 -9.36 -12.09 Z2750653629 2.73 -11.26 1.37 0.83 1.90 1.90 \nZ4767467072 N alpha2 0.01 0.10 -9.53 -12.09 Z2750653629 2.56 -12.64 2.78 -0.55 3.11 3.11 \nZ4954471231 C alpha2 5.48 0.00 -13.10 -12.09 Z2750653629 -1.01 -12.58 1.92 -0.49 0.52 0.52 \nZ5295862251 N alpha2 0.09 0.01 -10.69 -12.09 Z2750653629 1.40 -13.12 1.85 -1.03 2.43 2.43 \nZ5874039302 N alpha2 0.01 0.14 -9.34 -12.09 Z2750653629 2.75 -11.09 2.73 1.00 1.75 1.75 \nZ8598084898 N alpha2 0.01 0.16 -9.29 -12.09 Z2750653629 2.80 -12.08 2.00 0.01 2.79 2.79 \nZ8598084899 C alpha2 0.01 0.20 -9.13 -12.09 Z2750653629 2.96 -12.34 2.13 -0.25 3.21 3.21 \nZ8598084900 C alpha2 0.36 0.00 -11.49 -12.09 Z2750653629 0.60 -12.09 1.70 0.00 0.60 0.60 \nZ8598084901 C alpha2 9.72 0.00 -13.44 -12.09 Z2750653629 -1.35 -11.71 2.51 0.38 1.73 1.73 \nZ8598084905 C alpha2 52.17 0.00 -14.43 -12.09 Z2750653629 -2.34 -12.60 1.46 -0.51 1.83 1.83 \nZ8598084906 C alpha2 10.68 0.00 -13.49 -12.09 Z2750653629 -1.40 -12.37 1.86 -0.28 1.12 1.12 \nZ8598084908 F alpha2 0.16 0.01 -11.02 -12.09 Z2750653629 1.07 -12.62 1.57 -0.53 1.60 1.60 \nZ8598084909 F alpha2 2.47 0.00 -12.63 -12.09 Z2750653629 -0.54 -13.60 1.86 -1.51 0.97 0.97 \nZ8598084911 F alpha2 1.70 0.00 -12.40 -12.09 Z2750653629 -0.31 -11.76 1.85 0.33 0.64 0.64 \nZ8598084912 Cl alpha2 0.10 0.01 -10.71 -12.09 Z2750653629 1.37 -12.93 1.52 -0.84 2.21 2.21 \nZ8598084913 Cl alpha2 1.56 0.00 -12.35 -12.09 Z2750653629 -0.26 -12.77 2.04 -0.67 0.41 0.41 \nZ8598084914 Cl alpha2 0.08 0.02 -10.61 -12.09 Z2750653629 1.48 -13.39 1.82 -1.30 2.77 2.77 \nZ8598084918 Cl alpha2 2.38 0.00 -12.60 -12.09 Z2750653629 -0.51 -11.61 1.86 0.48 0.99 0.99 \nZ8598084922 O alpha2 0.67 0.00 -11.85 -12.09 Z2750653629 0.24 -13.02 1.90 -0.93 1.17 1.17 \nZ8598084925 O alpha2 0.18 0.01 -11.07 -12.09 Z2750653629 1.02 -8.13 1.45 3.96 2.95 2.95 \nZ8701729806 O alpha2 4.80 0.00 -13.02 -12.09 Z2750653629 -0.93 -12.59 2.37 -0.50 0.43 0.43 \nZ8731865364 O alpha2 0.17 0.01 -11.06 -12.09 Z2750653629 1.03 -12.14 2.06 -0.05 1.09 1.09 \nZ8747421921 O alpha2 1.51 0.00 -12.33 -12.09 Z2750653629 -0.24 -12.16 2.30 -0.07 0.17 0.18 \nZ8755839162  O alpha2 2.82 0.00 -12.70 -12.09 Z2750653629 -0.61 -12.60 2.42 -0.51 0.10 0.10 \nZ8825803704 C alpha2 0.58 0.00 -11.77 -12.09 Z2750653629 0.32 -11.65 2.44 0.44 0.11 0.11 \nZ8825803705 Cl alpha2 0.00 1.94 -7.79 -12.09 Z2750653629 4.30 -8.73 1.28 3.36 0.93 0.93 \nZ3034773353 C alpha2 0.31 72.29 -5.65 -6.35 Z3034773248 0.70 -5.49 1.33 0.86 0.16 0.16 \nZ3034773362 C alpha2 0.00 8270.00 -2.84 -6.35 Z3034773248 3.51 -5.73 1.31 0.62 2.89 2.89 \nZ3034773379 Cl alpha2 0.98 22.58 -6.34 -6.35 Z3034773248 0.01 -6.08 1.51 0.27 0.26 0.26 \nZ3034773381 Cl alpha2 0.49 45.57 -5.92 -6.35 Z3034773248 0.43 -6.25 1.40 0.10 0.33 0.33 \nZ3034773663 N alpha2 6.95 3.18 -7.50 -6.35 Z3034773248 -1.15 -5.82 1.77 0.53 1.68 1.68 \nZ3305295413 F alpha2 0.71 31.07 -6.15 -6.35 Z3034773248 0.20 -6.26 1.21 0.09 0.11 0.11 \nZ8727395864 O alpha2 2.38 9.29 -6.86 -6.35 Z3034773248 -0.51 -6.03 2.42 0.32 0.83 0.83 \nZ8836317802 O alpha2 0.35 63.92 -5.72 -6.35 Z3034773248 0.63 -5.77 1.88 0.59 0.04 0.04 \nZ8904328913 F alpha2 1.69 13.12 -6.66 -6.35 Z3034773248 -0.31 -6.07 1.42 0.28 0.59 0.59 \nZ8904332023 O alpha2 1.55 14.31 -6.61 -6.35 Z3034773248 -0.26 -4.34 1.69 2.02 2.27 2.27 \nZ8908003437 O alpha2 8.45 2.62 -7.61 -6.35 Z3034773248 -1.26 -6.38 1.72 -0.02 1.24 1.24 \n Z3034773358 C alpha2 33.63 0.66 -8.43 -6.35 Z3034773248 -2.08 -7.08 1.55 -0.73 1.36 1.36 \nZ6071720064 C alpha2 7.06 0.00 -12.18 -11.02 Z4376630014 -1.16 -11.03 1.21 -0.00 1.15 1.15 \nZ6071720076 Cl alpha2 0.15 0.06 -9.88 -11.02 Z4376630014 1.14 -9.98 1.15 1.04 0.10 0.10 \nZ6071720077 F alpha2 0.07 0.12 -9.45 -11.02 Z4376630014 1.58 -11.48 0.90 -0.46 2.03 2.03 \nZ6071720079 F alpha2 2.70 0.00 -11.61 -11.02 Z4376630014 -0.59 -11.51 1.33 -0.49 0.10 0.10 \nZ6071720081 Cl alpha2 0.02 0.38 -8.75 -11.02 Z4376630014 2.27 -11.62 0.89 -0.59 2.87 2.87 \nZ8081978921 Cl alpha2 0.00 28.99 -6.19 -11.02 Z4376630014 4.83 -12.87 0.94 -1.85 6.68 6.68 \nZ8598084957 C alpha2 0.04 0.21 -9.11 -11.02 Z4376630014 1.91 -10.84 1.25 0.18 1.73 1.73 \nZ8598084962 Cl alpha2 0.00 17.51 -6.49 -11.02 Z4376630014 4.53 -13.05 1.15 -2.03 6.57 6.56 \nZ3034773369 F alpha2 2.28 5.85 -7.14 -6.65 Z8727395870 -0.49 -6.18 1.23 0.47 0.96 0.96 \nZ3034773371 F alpha2 0.78 17.08 -6.50 -6.65 Z8727395870 0.15 -6.41 1.28 0.24 0.09 0.09 \nZ8727394867 Cl alpha2 1.09 12.23 -6.70 -6.65 Z8727395870 -0.05 -5.95 1.27 0.70 0.75 0.75 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 58 \nName modification protein affinity/potency- \nfold \nAffinity/potenc\ny_uM \nexp_G Parent exp \nG \nparent \nname \nexp_G FEP_G fep_unc fep_G fep_G \nerr_abs \nfep_G \nerr_abs \nZ8731642722 F alpha2 1.32 10.10 -6.81 -6.65 Z8727395870 -0.16 -5.51 1.64 1.14 1.30 1.30 \nZ8731733054 C alpha2 1.81 7.36 -7.00 -6.65 Z8727395870 -0.35 -5.67 1.39 0.98 1.33 1.33 \nZ8735760397 C alpha2 0.95 13.97 -6.62 -6.65 Z8727395870 0.03 -6.27 1.50 0.38 0.35 0.35 \nZ8836317766 C alpha2 0.90 14.78 -6.59 -6.65 Z8727395870 0.06 -6.71 1.36 -0.06 0.12 0.12 \nZ8836317772 O alpha2 0.40 33.53 -6.10 -6.65 Z8727395870 0.55 -6.41 2.40 0.24 0.31 0.31 \nZ8904328914 C alpha2 0.66 20.12 -6.41 -6.65 Z8727395870 0.24 -6.06 1.46 0.59 0.34 0.35 \nZ8904332000 Cl alpha2 3.68 3.62 -7.42 -6.65 Z8727395870 -0.77 -6.81 1.57 -0.16 0.61 0.61 \nZ8904332024 O alpha2 0.00 24700.00 -2.19 -6.65 Z8727395870 4.46 -6.77 2.74 -0.12 4.58 4.57 \nZ8904332025 O alpha2 0.85 15.66 -6.55 -6.65 Z8727395870 0.10 -7.01 2.66 -0.36 0.45 0.45 \nZ8908003436 O alpha2 0.65 20.46 -6.40 -6.65 Z8727395870 0.25 -6.07 2.55 0.58 0.33 0.33 \nZ9089437746 Cl alpha2 0.72 18.62 -6.45 -6.65 Z8727395870 0.20 -7.08 1.27 -0.43 0.63 0.63 \nZ2610512538 Cl ampc 1.37 17.60 -6.49 -6.30 Z2610488449 -0.19 -6.31 0.43 0.18 0.18 0.37 \nZ2610513554 F ampc 0.95 25.46 -6.27 -6.30 Z2610488449 0.03 -7.03 0.52 -0.54 0.76 0.57 \nZ2610513975 F ampc 0.74 32.65 -6.12 -6.30 Z2610488449 0.18 -6.50 0.50 -0.01 0.38 0.19 \nZ8584748090 F ampc 1.41 17.13 -6.50 -6.30 Z2610488449 -0.20 -6.49 0.41 0.00 0.01 0.20 \nZ8584748099 N ampc 8.38 2.89 -7.56 -6.30 Z2610488449 -1.26 -6.61 0.41 -0.12 0.95 1.14 \nZ8584748244 F ampc 1.68 14.42 -6.60 -6.30 Z2610488449 -0.31 -6.39 0.41 0.10 0.21 0.41 \nZ8584748322 Cl ampc 1.22 19.77 -6.42 -6.30 Z2610488449 -0.12 -6.99 0.43 -0.50 0.57 0.38 \nZ8584748346 Cl ampc 2.72 8.88 -6.89 -6.30 Z2610488449 -0.59 -6.83 0.41 -0.34 0.06 0.25 \nZ1355328651 Cl cb2 36.17 0.06 -9.81 -7.69 Z52076138 -2.13 -9.45 0.60 -1.85 0.36 0.27 \nZ1511587581 C cb2 2.66 0.87 -8.26 -7.69 Z52076138 -0.58 -8.29 0.52 -0.69 0.03 0.11 \nZ1745610848 C cb2 21.10 0.11 -9.49 -7.69 Z52076138 -1.81 -7.94 0.41 -0.34 1.55 1.46 \nZ1820897159 F cb2 9.74 0.24 -9.03 -7.69 Z52076138 -1.35 -7.91 0.41 -0.31 1.12 1.04 \nZ185279188 C cb2 9.83 0.24 -9.04 -7.69 Z52076138 -1.35 -8.22 0.49 -0.62 0.82 0.73 \nZ6755334561 Cl cb2 17.44 0.13 -9.38 -7.69 Z52076138 -1.69 -8.91 0.49 -1.31 0.47 0.38 \nZ8383080706 C cb2 19.26 0.12 -9.44 -7.69 Z52076138 -1.75 -9.50 0.45 -1.90 0.06 0.15 \nZ8701681958 C cb2 0.04 11.24 -6.75 -8.72 Z6969215903 1.97 -6.62 0.91 1.33 0.13 0.65 \nZ8701681960 C cb2 0.07 5.53 -7.17 -8.72 Z6969215903 1.55 -6.83 0.84 1.12 0.34 0.44 \nZ8701681963 C cb2 0.49 0.83 -8.30 -8.72 Z6969215903 0.43 -8.15 0.43 -0.20 0.15 0.63 \nZ8701681979 F cb2 0.06 7.23 -7.01 -8.72 Z6969215903 1.71 -8.18 0.83 -0.23 1.17 1.95 \nZ8701681983 Cl cb2 1.09 0.37 -8.77 -8.72 Z6969215903 -0.05 -8.29 0.43 -0.34 0.48 0.30 \nZ8701681984 Cl cb2 0.11 3.74 -7.40 -8.72 Z6969215903 1.32 -7.51 0.91 0.44 0.11 0.89 \nZ8701681647 N cb2 0.14 0.90 -8.24 -9.42 Z8184698918 1.18 -10.25 1.15 -0.84 2.01 2.02 \nZ8701681649 C cb2 1.80 0.07 -9.77 -9.42 Z8184698918 -0.35 -7.41 0.47 2.00 2.36 2.35 \nZ8701681656 C cb2 29.61 0.00 -11.43 -9.42 Z8184698918 -2.01 -11.20 0.98 -1.79 0.23 0.22 \nZ8701681673 O cb2 0.20 0.62 -8.46 -9.42 Z8184698918 0.96 -4.54 1.25 4.87 3.92 3.91 \nZ8701681677 F cb2 1.57 0.08 -9.69 -9.42 Z8184698918 -0.27 -8.12 0.47 1.29 1.57 1.56 \nZ8701681695 Cl cb2 58.46 0.00 -11.83 -9.42 Z8184698918 -2.41 -11.01 1.25 -1.60 0.82 0.81 \nZ8701681697 Cl cb2 1.40 0.09 -9.62 -9.42 Z8184698918 -0.20 -7.71 0.46 1.70 1.91 1.90 \nZ9187344338 N mac1 0.09 7.96 -6.96 -8.40 Z1039063794 1.44 -7.10 0.43 1.30 0.14 0.14 \nZ9187344350 N mac1 0.00 935.70 -4.13 -8.40 Z1039063794 4.26 -5.56 0.60 2.83 1.43 1.43 \nZ9187344358 N mac1 0.46 1.54 -7.93 -8.40 Z1039063794 0.47 -8.99 0.38 -0.59 1.06 1.06 \nZ1359487111 C mac1 0.00 745.80 -4.27 -8.40 Z1039063794 4.13 -6.41 0.43 1.99 2.14 2.14 \nZ8756044544 C mac1 0.10 6.94 -7.04 -8.40 Z1039063794 1.36 -7.82 0.48 0.57 0.78 0.78 \nZ8756044539 C mac1 0.16 4.30 -7.32 -8.40 Z1039063794 1.08 -8.69 0.44 -0.29 1.37 1.37 \nZ237580338 N mac1 0.06 11.59 -6.73 -8.40 Z1039063794 1.66 -8.24 0.76 0.15 1.51 1.51 \nZ9043825450 O mac1 0.04 19.29 -6.43 -8.40 Z1039063794 1.96 -3.50 0.39 4.90 2.93 2.93 \nZ4223060659 N mac1 0.04 18.18 -6.47 -8.40 Z1039063794 1.93 -9.73 0.34 -1.33 3.26 3.26 \nZ8207543137 Cl mac1 0.00 220.00 -4.99 -8.40 Z1039063794 3.41 -6.79 0.46 1.60 1.80 1.80 \nZ8207599372 Cl mac1 0.06 12.22 -6.70 -8.40 Z1039063794 1.69 -9.59 0.45 -1.19 2.88 2.88 \nZ8207316857 Cl mac1 0.05 13.37 -6.65 -8.40 Z1039063794 1.75 -8.70 0.36 -0.30 2.05 2.05 \nZ8598075280 N mac1 0.16 15.18 -6.57 -7.72 Z5398393122 1.14 -7.94 0.52 -0.23 1.37 1.37 \nZ7691912366 C mac1 0.77 3.23 -7.49 -7.72 Z5398393122 0.23 -7.31 0.62 0.41 0.18 0.18 \nZ8598075283 C mac1 0.47 5.29 -7.20 -7.72 Z5398393122 0.52 -8.78 0.56 -1.06 1.58 1.58 \nZ8598075292 Cl mac1 2.24 1.12 -8.12 -7.72 Z5398393122 -0.40 -7.38 0.56 0.34 0.74 0.74 \nZ8598075284 C mac1 0.40 6.25 -7.10 -7.72 Z5398393122 0.62 -7.65 0.62 0.07 0.55 0.55 \nZ8928826043 C mac1 0.36 27.75 -6.22 -6.89 Z7534253453 0.68 -6.49 0.48 0.40 0.28 0.28 \nZ8727396582 C mac1 0.70 14.15 -6.61 -6.89 Z7534253453 0.28 -7.02 0.41 -0.13 0.41 0.41 \nZ8727396638 C mac1 4.13 2.40 -7.67 -6.89 Z7534253453 -0.77 -6.54 0.61 0.35 1.12 1.12 \nZ8929429249 C mac1 0.19 52.56 -5.84 -6.89 Z7534253453 1.05 -8.03 0.50 -1.14 2.19 2.19 \nZ8990523176 C mac1 0.46 21.65 -6.36 -6.89 Z7534253453 0.53 -7.28 0.55 -0.39 0.92 0.92 \nZ8990523178 O mac1 39.68 0.25 -9.01 -6.89 Z7534253453 -2.11 -7.14 0.65 -0.25 1.86 1.86 \nZ8990919048 O mac1 0.32 30.69 -6.16 -6.89 Z7534253453 0.74 -7.45 0.64 -0.56 1.30 1.30 \nZ8598075302 O mac1 0.70 14.27 -6.61 -6.89 Z7534253453 0.28 -6.04 0.60 0.85 0.57 0.57 \nZ8990523184 O mac1 19.84 0.50 -8.60 -6.89 Z7534253453 -1.70 -8.49 0.55 -1.60 0.10 0.10 \nZ7692056404 F mac1 1.79 5.53 -7.17 -6.89 Z7534253453 -0.28 -7.45 0.42 -0.56 0.28 0.28 \nZ8928826033 F mac1 2.82 3.51 -7.44 -6.89 Z7534253453 -0.55 -8.56 0.49 -1.67 1.12 1.12 \nZ8990523169 N mac1 0.99 10.04 -6.82 -6.89 Z7534253453 0.07 -7.01 0.61 -0.12 0.19 0.19 \nZ8990527720 F mac1 0.94 10.51 -6.79 -6.89 Z7534253453 0.10 -6.49 0.47 0.40 0.30 0.30 \nZ7692046343 Cl mac1 4.75 2.09 -7.75 -6.89 Z7534253453 -0.86 -7.62 0.56 -0.73 0.13 0.13 \nZ8768700676 Cl mac1 0.46 21.48 -6.37 -6.89 Z7534253453 0.52 -5.15 0.79 1.74 1.22 1.22 \nZ8990523194 Cl mac1 6.04 1.64 -7.89 -6.89 Z7534253453 -1.00 -7.27 0.69 -0.38 0.62 0.62 \nZ8727401304 Cl mac1 5.17 1.92 -7.80 -6.89 Z7534253453 -0.91 -8.74 0.52 -1.85 0.94 0.94 \nZ7692056627 N mac1 1.14 8.67 -6.91 -6.89 Z7534253453 -0.01 -6.90 0.72 -0.01 0.01 0.01 \nZ8929428675 N mac1 2.22 4.46 -7.30 -6.89 Z7534253453 -0.41 -7.99 0.52 -1.10 0.69 0.69 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 59 \nName modification protein affinity/potency- \nfold \nAffinity/potenc\ny_uM \nexp_G Parent exp \nG \nparent \nname \nexp_G FEP_G fep_unc fep_G fep_G \nerr_abs \nfep_G \nerr_abs \nZ7691912473 N mac1 0.30 33.10 -6.11 -6.89 Z7534253453 0.78 -6.03 0.50 0.86 0.08 0.08 \nZ7140729870 C mac1 1.30 7.61 -6.98 -6.89 Z7534253453 -0.09 -6.12 0.41 0.77 0.86 0.86 \nZ3033185506 N sert 0.07 16.52 -6.52 -8.13 Z2009978218 1.61 -7.24 2.99 0.89 0.72 0.72 \nZ3991412496 F sert 0.93 1.18 -8.09 -8.13 Z2009978218 0.04 -8.86 2.18 -0.73 0.77 0.78 \nZ4224292450 O sert 0.20 5.53 -7.17 -8.13 Z2009978218 0.96 -5.66 2.48 2.47 1.51 1.51 \nZ4767467074 Cl sert 9.48 0.12 -9.46 -8.13 Z2009978218 -1.33 -9.28 2.39 -1.16 0.18 0.18 \nZ4767467093 C sert 1.22 0.89 -8.25 -8.13 Z2009978218 -0.12 -8.71 2.60 -0.59 0.46 0.47 \nZ4955958409 N sert 0.12 9.07 -6.88 -8.13 Z2009978218 1.25 -4.95 2.39 3.17 1.92 1.92 \nZ4957966902 N sert 0.02 60.37 -5.76 -8.13 Z2009978218 2.38 -6.02 2.68 2.11 0.26 0.27 \nZ5022589075 C sert 4.16 0.26 -8.98 -8.13 Z2009978218 -0.84 -8.81 2.46 -0.68 0.17 0.16 \nZ5425693717 Cl sert 0.44 2.49 -7.64 -8.13 Z2009978218 0.49 -7.60 2.42 0.53 0.04 0.04 \nZ5766294474 F sert 1.04 1.05 -8.16 -8.13 Z2009978218 -0.02 -8.04 2.44 0.08 0.11 0.11 \nZ8646465742 Cl sert 5.05 0.22 -9.09 -8.13 Z2009978218 -0.96 -9.41 2.57 -1.29 0.32 0.33 \nZ8827560439 C sert 2.92 0.37 -8.77 -8.13 Z2009978218 -0.63 -9.59 2.68 -1.47 0.83 0.83 \nZ8827560440 C sert 0.75 1.46 -7.96 -8.13 Z2009978218 0.17 -8.73 2.30 -0.60 0.76 0.77 \nZ8827560451 O sert 0.14 7.56 -6.99 -8.13 Z2009978218 1.15 -5.28 2.99 2.84 1.70 1.70 \nZ2573292509 C sert 0.35 0.24 -9.04 -9.67 Z2573292480 0.63 -12.05 1.66 -2.38 3.01 3.01 \nZ3498174416 N sert 0.03 2.64 -7.61 -9.67 Z2573292480 2.06 -7.12 1.18 2.56 0.49 0.50 \nZ4067645979 N sert 0.03 2.98 -7.54 -9.67 Z2573292480 2.13 -8.58 1.19 1.09 1.04 1.04 \nZ4224285938 N sert 0.14 0.58 -8.50 -9.67 Z2573292480 1.16 -9.97 1.47 -0.30 1.47 1.46 \nZ4765367699 C sert 0.69 0.12 -9.45 -9.67 Z2573292480 0.22 -10.84 1.59 -1.17 1.39 1.38 \nZ5458682006 C sert 3.17 0.03 -10.35 -9.67 Z2573292480 -0.68 -9.29 1.18 0.39 1.06 1.07 \nZ8055284773 N sert 0.50 0.16 -9.25 -9.67 Z2573292480 0.41 -10.21 1.26 -0.54 0.96 0.95 \nZ8501235596 C sert 1.27 0.06 -9.81 -9.67 Z2573292480 -0.14 -9.79 1.91 -0.11 0.02 0.03 \nZ8647047828 C sert 1.54 0.05 -9.92 -9.67 Z2573292480 -0.26 -7.84 1.13 1.83 2.09 2.09 \nZ8827560409 O sert 0.24 0.35 -8.81 -9.67 Z2573292480 0.85 -9.20 1.55 0.47 0.39 0.38 \nZ8827560428 Cl sert 1.56 0.05 -9.93 -9.67 Z2573292480 -0.26 -10.67 1.27 -0.99 0.74 0.73 \nZ3950610525 N sert 0.26 0.91 -8.24 -9.04 Z2573292509 0.80 -9.77 1.42 -0.70 1.53 1.49 \nZ4067646017 N sert 0.02 10.81 -6.77 -9.04 Z2573292509 2.26 -8.04 1.61 1.03 1.27 1.23 \nZ4765367717 C sert 2.28 0.10 -9.53 -9.04 Z2573292509 -0.49 -10.88 1.45 -1.81 1.35 1.32 \nZ5106507813 N sert 0.01 16.52 -6.52 -9.04 Z2573292509 2.52 -7.13 1.56 1.94 0.61 0.57 \nZ5458682044 C sert 16.14 0.01 -10.69 -9.04 Z2573292509 -1.65 -9.21 1.48 -0.13 1.47 1.51 \nZ8647047832 C sert 1.79 0.13 -9.39 -9.04 Z2573292509 -0.35 -7.97 1.44 1.11 1.42 1.46 \nZ8781161134 N sert 0.07 3.57 -7.43 -9.04 Z2573292509 1.61 -7.91 1.74 1.17 0.48 0.44 \nZ8827560496 O sert 0.63 0.37 -8.77 -9.04 Z2573292509 0.27 -9.10 1.55 -0.02 0.33 0.29 \nZ8827560515 Cl sert 2.76 0.09 -9.64 -9.04 Z2573292509 -0.60 -9.58 1.85 -0.50 0.06 0.10 \nZ6971277405 N sert 0.55 24.20 -6.30 -6.65 Z6971277399 0.35 -5.25 1.28 1.23 1.04 0.88 \nZ6971278430 C sert 24.81 0.54 -8.55 -6.65 Z6971277399 -1.90 -6.44 1.28 0.03 2.11 1.94 \nZ6971278455 C sert 4.08 3.28 -7.48 -6.65 Z6971277399 -0.83 -6.21 1.35 0.27 1.27 1.10 \nZ6971278464 C sert 3.99 3.36 -7.47 -6.65 Z6971277399 -0.82 -6.23 1.36 0.24 1.23 1.06 \nZ6971278779 F sert 0.42 31.49 -6.14 -6.65 Z6971277399 0.51 -6.54 1.43 -0.06 0.39 0.56 \nZ6971279383 Cl sert 5.09 2.62 -7.61 -6.65 Z6971277399 -0.96 -6.97 1.28 -0.49 0.64 0.47 \nZ6971279394 Cl sert 19.99 0.67 -8.42 -6.65 Z6971277399 -1.77 -7.37 1.29 -0.89 1.05 0.89 \nZ6971279399 Cl sert 3.57 3.74 -7.40 -6.65 Z6971277399 -0.75 -7.00 1.25 -0.52 0.40 0.23 \nZ6971279682 Br sert 11.93 1.12 -8.12 -6.65 Z6971277399 -1.47 -7.71 1.44 -1.23 0.40 0.24 \nZ6971279689 Br sert 7.58 1.76 -7.85 -6.65 Z6971277399 -1.20 -7.15 1.59 -0.68 0.69 0.52 \nZ6971279702 Br sert 16.09 0.83 -8.29 -6.65 Z6971277399 -1.65 -7.25 1.59 -0.77 1.04 0.88 \nZ8172508405 O sert 4.18 3.20 -7.50 -6.65 Z6971277399 -0.85 -6.87 1.48 -0.39 0.63 0.46 \nZ3034773248 Cl sert 9.24 4.62 -7.28 -5.96 Z8727393896 -1.32 -6.68 0.99 -1.00 0.60 0.32 \nZ8024305237 Br sert 9.88 4.32 -7.32 -5.96 Z8727393896 -1.36 -6.48 1.17 -0.79 0.84 0.56 \nZ8727394028 O sert 5.45 7.84 -6.96 -5.96 Z8727393896 -1.00 -6.79 1.32 -1.11 0.17 0.11 \nZ8727394030 Br sert 7.56 5.65 -7.16 -5.96 Z8727393896 -1.20 -6.76 0.96 -1.08 0.40 0.12 \nZ8727395866 F sert 2.44 17.51 -6.49 -5.96 Z8727393896 -0.53 -6.26 1.02 -0.57 0.23 0.05 \nZ8727395870 Cl sert 5.79 7.37 -7.00 -5.96 Z8727393896 -1.04 -6.20 1.00 -0.52 0.80 0.52 \nZ8727396158 Br sert 10.50 4.07 -7.35 -5.96 Z8727393896 -1.39 -6.74 1.07 -1.05 0.62 0.34 \nZ8731642686 C sert 23.33 1.83 -7.83 -5.96 Z8727393896 -1.87 -6.64 1.00 -0.95 1.19 0.91 \nZ8735711409 O sert 1.04 40.96 -5.99 -5.96 Z8727393896 -0.02 -5.19 1.42 0.49 0.79 0.52 \nZ8735711431 Cl sert 6.03 7.08 -7.03 -5.96 Z8727393896 -1.06 -6.20 0.98 -0.52 0.82 0.54 \nZ3034773283 C sert 20.47 0.09 -9.61 -7.83 Z8731642686 -1.79 -9.35 1.32 -1.51 0.26 0.28 \nZ3034773290 C sert 1.75 1.05 -8.16 -7.83 Z8731642686 -0.33 -8.86 0.97 -1.02 0.70 0.69 \nZ3034773359 Cl sert 4.14 0.44 -8.67 -7.83 Z8731642686 -0.84 -8.49 1.22 -0.65 0.17 0.19 \nZ3034773386 Br sert 27.20 0.07 -9.78 -7.83 Z8731642686 -1.96 -9.45 1.29 -1.61 0.33 0.35 \nZ8727393902 Br sert 7.28 0.25 -9.00 -7.83 Z8731642686 -1.18 -8.75 1.23 -0.91 0.25 0.27 \nZ8741436112 F sert 6.98 0.26 -8.98 -7.83 Z8731642686 -1.15 -7.57 1.57 0.27 1.41 1.42 \nZ8836317766 Cl sert 5.87 0.31 -8.88 -7.83 Z8731642686 -1.05 -9.03 1.18 -1.18 0.15 0.13 \nZ8843617521 O sert 0.17 10.72 -6.78 -7.83 Z8731642686 1.05 -7.17 1.44 0.68 0.39 0.37 \nZ8843617522 O sert 0.39 4.71 -7.27 -7.83 Z8731642686 0.56 -8.15 1.60 -0.31 0.89 0.87 \nZ8843617525 O sert 0.53 3.46 -7.45 -7.83 Z8731642686 0.38 -9.65 1.36 -1.80 2.20 2.18 \nZ8890819769 F sert 0.35 5.27 -7.20 -7.83 Z8731642686 0.63 -7.55 1.50 0.29 0.35 0.33 \nZ8891248505 C sert 3.56 0.51 -8.58 -7.83 Z8731642686 -0.75 -7.44 1.14 0.41 1.14 1.16 \nZ8904328909 N sert 0.20 9.05 -6.88 -7.83 Z8731642686 0.95 -5.86 1.27 1.99 1.02 1.04 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 60 \nName modification protein affinity/potency- \nfold \nAffinity/potenc\ny_uM \nexp_G Parent exp \nG \nparent \nname \nexp_G FEP_G fep_unc fep_G fep_G \nerr_abs \nfep_G \nerr_abs \nZ8904328910 F sert 2.19 0.84 -8.29 -7.83 Z8731642686 -0.46 -8.82 1.12 -0.97 0.53 0.51 \nZ8904332022 O sert 11.48 0.16 -9.27 -7.83 Z8731642686 -1.45 -8.01 1.60 -0.17 1.26 1.28 \nZ8915636241 C sert 4.56 0.40 -8.72 -7.83 Z8731642686 -0.90 -8.44 1.09 -0.59 0.29 0.31 \n Z3034773358 Cl sert 23.52 0.08 -9.70 -7.83 Z8731642686 -1.87 -9.38 1.27 -1.54 0.32 0.33 \na, ΔGexp (kcal mol⁻¹): Experimental free energy calculated from assay-derived K (in M) as ΔGexp = 1168 \nRT ln K, with R = 1.987 × 10⁻³ kcal mol⁻¹ K⁻¹ and T = 298.15 K. 1169 \nb, ΔGexp,parent (kcal mol⁻¹): ΔGexp of the corresponding parent compound for the same target. 1170 \nc, ΔΔGexp (kcal mol⁻¹): Relative experimental free energy versus parent, ΔΔGexp = ΔGexp,analog − 1171 \nΔGexp,parent. 1172 \nd, ΔGFEP (kcal mol⁻¹): Absolute binding free energy predicted by FEP+. 1173 \ne, σFEP (kcal mol⁻¹): Uncertainty reported by the FEP+ workflow for ΔGFEP. 1174 \nf, ΔΔGFEP (kcal mol⁻¹): Relative FEP+ free energy versus parent, ΔΔGFEP = ΔGFEP,analog − 1175 \nΔGFEP,parent. 1176 \ng, |ΔGFEP − ΔGexp| (kcal mol⁻¹): Absolute error in ΔG. 1177 \nh, |ΔΔGFEP − ΔΔGexp| (kcal mol⁻¹): Absolute error in ΔΔG.Positive ΔΔG\\Delta\\Delta GΔΔG 1178 \nindicates decreased potency/affinity relative to the parent; negative ΔΔG\\Delta\\Delta GΔΔG indicates 1179 \nincreased potency/affinity. Blank entries indicate values not available for the corresponding metric. 1180 \nFull chemical structures (including SMILES) for all parent compounds and analogs are provided in 1181 \nSupplementary Table 1. 1182 \n 1183 \n 1184 \n 1185 \n 1186 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n1187 \n 1188 \nExtended Data Table 2. The effect of the analogs, relative to parents, on in vitro pharmacokinetics* 1189 \nparent target Total \ncompound\ns \nMicrosome stability \nT1/2, min fold change \n(≥3x) \nPlasma stability \nT1/2, min fold \nchange (≥3x) \nppb % of unbound \ncompound fold \nchange (≥3x) \nSolubility, µM \nfold change \n(≥3x) \nhERG \napproximate \nIC50 µM fold \nchange (≥3x) \npermeabilit\ny cm/s fold \nchange \n(≥3x) \nZ4407498716 MOR 16 0 0 0 0 0 0 \nZ2750653629 alpha\n2 \n31 7 2 4 5 5 0 \nZ3034773248 alpha\n2 \n18 1 1 3 16 7 0 \nZ4376630014 alpha\n2 \n10 5 1 7 2 1 0 \nZ8727395870 alpha\n2 \n18 3 0 6 16 11 0 \nZ2610488449 ampc 10 1 0 2 0 2 4 \nZ52076138 cb2 7 0 1 0 1 0 0 \nZ6969215903 cb2 6 1 2 2 0 0 0 \nZ818469891\n8 \ncb2 20 1 4 0 0 5 0 \nZ1039063794 mac1 12 4 0 1 0 3 0 \nZ5398393122 mac1 5 4 1 0 0 0 4 \nZ7534253453 mac1 22 1 10 0 0 4 9 \nZ2009978218 sert 14 5 4 3 0 2 0 \nZ2573292480 sert 11 3 0 3 0 0 0 \nZ2573292509 sert 9 3 2 4 6 3 0 \nZ6971277399 sert 15 1 3 1 0 1 0 \nZ8727393896 sert 14 0 3 0 0 0 0 \nZ8731642686 sert 19 0 3 3 0 1 0 \n* The MOR parent series and its 16 analogs are not included in this table because they were not profiled 1190 \nfor in vitro PK properties. 1191 \n 1192 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint \n\n 1193 \n(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. \nThe copyright holder for this preprintthis version posted May 13, 2026. ; https://doi.org/10.64898/2026.05.10.724162doi: bioRxiv preprint","source_license":"Public-Domain","license_restricted":false}