Improving (Q)SAR predictions by examining bias in the selection of compounds for experimental testing

ABSTRACT

Existing data on structures and biological activities are limited and distributed unevenly across distinct molecular targets and chemical compounds. The question arises if these data represent an unbiased sample of the general population of chemical-biological interactions. To answer this question, we analyzed ChEMBL data for 87,583 molecules tested against 919 protein targets using supervised and unsupervised approaches. Hierarchical clustering of the Murcko frameworks generated using Chemistry Development Toolkit showed that the available data form a big diffuse cloud without apparent structure. In contrast hereto, PASS-based classifiers allowed prediction whether the compound had been tested against the particular molecular target, despite whether it was active or not. Thus, one may conclude that the selection of chemical compounds for testing against specific targets is biased, probably due to the influence of prior knowledge. We assessed the possibility to improve (Q)SAR predictions using this fact: PASS prediction of the interaction with the particular target for compounds predicted as tested against the target has significantly higher accuracy than for those predicted as untested (average ROC AUC are about 0.87 and 0.75, respectively). Thus, considering the existing bias in the data of the training set may increase the performance of virtual screening.     

Conformational space of tetrakis(2-naphthalenyl)ethene: a naphthologous AIE luminogen

The AIE luminogen tetrakis(2-naphthalenyl)ethene (2-NA ₄ E) was synthesized by Barton’s double extrusion diazo-thione coupling method from 2,2′-dinaphthyl thioketone and 2,2′-(diazomethylene)bisnaphthylene in 77 % yield. The structure of 2-NA ₄ E was confirmed by its ¹H NMR and ¹³C NMR spectra with full assignments. 2-NA ₄ E and its parent tetraphenylethene (Ph ₄ E) have been subjected to a comprehensive computational DFT study, in search of their conformational spaces. Seven conformers and two transition states of 2-NA ₄ E have been located. Four conformers and one transition state of Ph ₄ E have been located. The conformers of 2-NA ₄ E and Ph ₄ E are not overcrowded, as indicated by the contact distances in the fjord and cove regions. The relative free energies (ΔG ₂₉₈) of the six most stable conformers of 2-NA ₄ E are in the narrow range of 2.3 kJ/mol; they make comparable contributions (12–29 %) to the equilibrium mixture. The energy barriers for the diastereomerization D ₂-Z,Z,Z,Z $ \rightleftharpoons $ ? D ₂-E,E,E,E via the transition state C ₁-Z,E,E,Z and for the enantiomerization C ₂-Z,Z,E,E $ \rightleftharpoons $ ? C ₂-E,E,Z,Z via the transition state C _i -Z,E,Z,E are only 29.8 and 29.0 kJ/mol, respectively, indicating very rapid rates of diastereomerization and enantiomerization at room temperature. The values of naphthalenyl torsion angles and ethenic twist angles in 2-NA ₄ E are almost identical to those in the parent Ph ₄ E. The previously proposed “bulkiness” of the naphthalenyl substituents and the validity of the restriction of naphthalenyl rotation are challenged. The analysis of the AIE effect in 2-NA ₄ E should take into account the intermolecular homochiral and heterochiral interactions between the conformers.     

Theoretical consideration of bulk free-radical copolymerization with allowance for the preferential sorption of monomers into globular nanoreactors

The thermodynamics of a globular solution of binary copolymer in two-component solvent is theoretically considered. Classification of phase diagrams of such a solution is performed by their topological appearance. Based on the thermodynamic theory, binary bulk free-radical copolymerization under low conversions is examined in systems where macroradicals being in globular conformational state represent isolated nanoreactors. The theoretical results achieved permit an explanation of all qualitative peculiarities observed experimentally under copolymerization in such anomalous systems.     

Regioselective Friedel–Crafts deacylations of polycyclic aromatic ketones in the pyrene series

1,3,6,8-tetrabenzoylpyrene (1,3,6,8-Bz ₄ PY) and 1,3,6-tribenzoylpyrene (1,3,6-Bz ₃ PY) were synthesized and their crystal structures were determined. The Friedel–Crafts deacylations in PPA of 1,3,6,8-Bz ₄ PY (at 120–200 °C) and of 1,3,6-Bz ₃ PY (at 80–160 °C) have been studied. The mono-deacylation of 1,3,6-Bz ₃ PY was regioselective and led to three dibenzoylpyrenes in the following order of relative amounts: 1,8-Bz ₂ PY > 1,6-Bz ₂ PY > 1,3-Bz ₂ PY. 1,3,6,8-Bz ₄ PY was resistant to deacylation at 120–160 °C. The deacylations of 1,3,6,8-Bz ₄ PY at 200 °C gave the polycyclic aromatic ketone (PAK) 8H-dibenzo[def,qr]chrysen-8-one (DBCO) via an intramolecular Scholl reaction. Two plausible pathways of the Friedel–Crafts deacylation of 1,3,6,8-Bz ₄ PY to give DBCO are proposed. A density functional theory (DFT) B3LYP/6-311(d,p) computational study of the conformational spaces of 1,3,6-Bz ₃ PY and 1,3,6,8-Bz ₄ PY was performed. The estimated energy barriers of formation of dibenzoylpyrenes by deacylation of 1,3,6-Bz ₃ PY increase in the following order: 1,8-Bz ₂ PY < 1,3-Bz ₂ PY < 1,6-Bz ₂ PY. A mechanism of the Friedel–Crafts deacylation of 1,3,6-Bz ₃ PY in PPA via the respective O-protonated ketone and σ-complexes is presented.     

Nachweis und Bestimmung von Aluminium

Ohne Zusammenfassung     

Overcrowded naphthologs of mono-bridged tetraarylethylenes: analogs of bistricyclic aromatic enes

The naphthalogous mono-bridged tetraarylethylenes 9,9′-di-(1-naphthylmethylene)-9H-fluorene (5) and 9,9′-di-(1-naphthylmethylene)-9H-xanthene (6), analogs of bifluorenylidene (1) and bixanthenylidene (2), have been synthesized and their molecular and crystal structures have been determined. Ene 5 has been prepared by two alternative synthetic routes. The molecular structures of 5 and 6 show that each of these enes has very small twist around the central double bond, but the two naphthalene rings in both 5 and 6 are highly twisted. According to the NMR study, 5 and 6 in solution adopt conformations which are similar to those found by X-ray crystal structure analysis. The notable upfield shifts of H¹ and H⁸ (6.11 and 6.83 ppm, respectively) and H² and H⁷ (6.70 and 6.44 ppm, respectively) in 5 and 6 are due to the shielding caused by the nearly orthogonally twisted naphthalene rings. The B3LYP/6-31G(d) calculations of 5, 6, and their 2-naphthyl and phenyl analogs have been performed. In the 1-naphthyl series, the more efficient conjugation between the naphthyl substituents and the central C=C and the overcrowding due to the peri-hydrogen atoms lead to higher twists of the naphthyl groups and to lower twists of the central C=C. In the 2-naphthyl series, the opposite effects are noted. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.

Biphenalenylidene: the forgotten bistricyclic aromatic ene. A theoretical study

The Lawesson reagent and P(2)S(5) mediated reductive coupling of phenalenone (6) gives LPAH peropyrene (5) in 47% and 54% yields. The mechanism of the reaction involves the formation of phenalenethione (10), Z- and E-1,1'-biphenalenylidene (3), and 9 as intermediates. The electrocyclization reaction of Z-3 to 9, followed by aromatization, gives 5. The results of an ab initio and DFT study of 3 and 2,2'-biphenalenylidene (12) are reported. E-3 and Z-3 have a diradical character with twist angles of 44.8 degrees and 57.8 degrees (at UB3LYP/6-311G**). Delta E(++)(Tot) = 10.2 kJ/mol and Delta G(++)(298) = 10.6 kJ/mol for E-3 <==> Z-3 diastereomerization. These unusually low energy barriers are due to the ground-state diradical destabilization and the aromatic stabilization of the transition state TS-3. Triplet Z-3 is higher in energy than singlets E-3 and Z-3 by 10.4 and 3.1 kJ/mol. In the concealed non-Kekulé 12, singlet 12 is more stable than the triplet by 1.3 kJ/mol. Singlet 12 is more stable than singlet E-3 by 2.0 kJ/mol, and orthogonal singlet TS-12 is lower in energy than singlet TS-3 by 6.0 kJ/mol. The energy barriers for the hexatriene-cyclohexadiene electrocyclization Z-3 --> 9 are Delta E(++)(Tot) = 94.8 and Delta G(++)(298) = 98.3 kJ/mol (at (U)B3LYP/6-31G). The reaction occurs thermally in a conrotatory mode.     

Naphthologs of overcrowded bistricyclic aromatic enes: (E)-bisbenzo[a]fluorenylidene

(E)-11H-Bisbenzo[a]fluorenylidene (E-6) was synthesized by Barton’s double extrusion diazo-thione coupling method from 11H-benzo[a]fluoren-11-thione (11) and 11-diazo-11H-benzo[a]fluorene (13). The reaction is probably thermodynamically controlled; in the event that the less stable Z -6 is also formed, it would rapidly undergo Z → E diastereomerization to give E -6. The B3LYP/6-311G(d,p) calculated diastereomerization barrier for Z -6 → E -6 is ΔG ₂₉₈ = 57.0 kJ/mol (13.6 kcal/mol). The calculated equilibrium constant K _eq(E -6 → Z -6) = 92:8 (at 298 K) is indicative of a marked diastereoselectivity of the reaction leading to E -6. The structure of E-6 was established by ¹H-NMR and ¹³C-NMR spectroscopies and by X-ray analysis. PAE E-6 crystallizes in the monoclinic space group C2/c. The unit cell of the crystal structure E -6 contains eight molecules, arranged as four pairs of enantiomers. PAE E -6 adopts a twisted conformation with the pure twist of the central C¹¹=C^11′ bond ω = 39°. The dihedral angle ν in E -6 is 60.6°, which is significantly higher than the respective dihedral angle in PAEs Z -6, 2, E -7, Z -7, 14, and 15. The large syn-pyramidalization angles at C¹¹ and C^11′ (χ = 12.6° and 14.8°) of E-6 indicates the enhanced strain in the fjord regions of the molecule. The enhanced twist is primarily attributed to the double benzo[a]annelation of the bifluorenylidene moiety at the fjord regions. The B3LYP/6-311G(d,p) calculated structure of E -6 is in a very good agreement with the experimental X-ray structure. PAE E -6 adopts a twisted conformation in solution, with the downfield chemical shift of H¹/H^1′ (8.31 ppm); H¹⁰/H^10′ (δ = 7.20 ppm) and H⁹/H^9′ (δ = 6.86 ppm) in E -6 are positioned above the planes of the opposing naphthalene rings. PAEs E -6 and Z -6 are significantly higher in energy than their corresponding benzo[b]annelated isomers E -7 and Z -7.