Applying medicinal chemistry strategies to understand odorant discrimination

More than meets the AI: The possibilities and limits of machine learning in olfaction

Can machine learning crack the code in the nose? Over the past decade, studies tried to solve the relation between chemical structure and sensory quality with Big Data. These studies advanced computational models of the olfactory stimulus, utilizing artificial intelligence to mine for clear correlations between chemistry and psychophysics. Computational perspectives promised to solve the mystery of olfaction with more data and better data processing tools. None of them succeeded, however, and it matters as to why this is the case. This article argues that we should be deeply skeptical about the trend to black-box the sensory system's biology in our theories of perception. Instead, we need to ground both stimulus models and psychophysical data on real causal-mechanistic explanations of the olfactory system. The central question is: Would knowledge of biology lead to a better understanding of the stimulus in odor coding than the one utilized in current machine learning models? That is indeed the case. Recent studies about receptor behavior have revealed that the olfactory system operates by principles not captured in current stimulus-response models. This may require a fundamental revision of computational approaches to olfaction, including its psychological effects. To analyze the different research programs in olfaction, we draw on Lloyd's "Logic of Research Questions," a philosophical framework which assists scientists in explicating the reasoning, conceptual commitments, and problems of a modeling approach in question.

Odorant metabolism of the olfactory cleft mucus in idiopathic olfactory impairment patients and healthy volunteers

BACKGROUND

It remains unclear whether the metabolic activity of nasal mucus in the olfactory and respiratory areas is different. Moreover, age- and olfaction-related changes may affect metabolism.

METHODS

Hexanal, octanal, and 2-methylbutanal were selected for in vitro metabolism analysis and compared between the olfactory cleft and respiratory mucus of participants < 50-year-old with normal olfaction using gas chromatography mass spectrometry. The metabolic activity of hexanal in the olfactory cleft mucus was further compared between three groups, (1) normal olfaction, age < 50 years old, (2) normal olfaction, age ≥50 years old, and (3) idiopathic olfactory impairment. To characterize the enzyme(s) responsible for aldehyde reduction, we also tested if epalr22897estat and 3,5-dichlorosalicylic acid, types of reductase inhibitors, affect metabolism.

RESULTS

Conversion of aldehydes to their corresponding alcohols was observed in the olfactory cleft and respiratory mucus. The metabolic production of hexanol, octanol, and 2-methybutanol was significantly higher in the olfactory cleft mucus than in the respiratory mucus (p < 0.01). The metabolic conversion of hexanal to hexanol in the mucus of the idiopathic olfactory impairment group was significantly lower than that in the age-matched normal olfaction group. Excluding the nicotinamide adenine dinucleotide phosphate (NADPH) regenerating system from the reaction mixture inhibited metabolism. The addition of either epalr22897estat or 3,5-dichlorosalicylic acid did not inhibit this metabolic conversion.

CONCLUSIONS

The enzymatic metabolism of odorants in the olfactory cleft mucus is markedly higher than in the respiratory mucus and decreases in patients with idiopathic olfactory impairment.

Finding the Brain in the Nose

Olfaction is fundamentally distinct from other sensory modalities. Natural odor stimuli are complex mixtures of volatile chemicals that interact in the nose with a receptor array that, in rodents, is built from more than 1,000 unique receptors. These interactions dictate a peripheral olfactory code, which in the brain is transformed and reformatted as it is broadcast across a set of highly interconnected olfactory regions. Here we discuss the problems of characterizing peripheral population codes for olfactory stimuli, of inferring the specific functions of different higher olfactory areas given their extensive recurrence, and of ultimately understanding how odor representations are linked to perception and action. We argue that, despite the differences between olfaction and other sensory modalities, addressing these specific questions will reveal general principles underlying brain function.

Smell compounds classification using UMAP to increase knowledge of odors and molecular structures linkages

This study aims to highlight the relationships between the structure of smell compounds and their odors. For this purpose, heterogeneous data sources were screened, and 6038 odorant compounds and their known associated odors (162 odor notes) were compiled, each individual molecule being represented with a set of 1024 structural fingerprint. Several dimensional reduction techniques (PCA, MDS, t-SNE and UMAP) with two clustering methods (k-means and agglomerative hierarchical clustering AHC) were assessed based on the calculated fingerprints. The combination of UMAP with k-means and AHC methods allowed to obtain a good representativeness of odors by clusters, as well as the best visualization of the proximity of odorants on the basis of their molecular structures. The presence or absence of molecular substructures has been calculated on odorant in order to link chemical groups to odors. The results of this analysis bring out some associations for both the odor notes and the chemical structures of the molecules such as "woody" and "spicy" notes with allylic and bicyclic structures, "balsamic" notes with unsaturated rings, both "sulfurous" and "citrus" with aldehydes, alcohols, carboxylic acids, amines and sulfur compounds, and "oily", "fatty" and "fruity" characterized by esters and with long carbon chains. Overall, the use of UMAP associated to clustering is a promising method to suggest hypotheses on the odorant structure-odor relationships.

Imaging the living brain: An argument for ruthless reductionism from olfactory neurobiology

Should theories of "higher-level" cognitive effects originate in "lower-level" molecular mechanisms? This paper supports reductionist explanations of sensory perception via molecular mechanisms in neurobiology. It shows that molecular and cellular mechanisms must constitute the material foundation to derive better theories and models for neuroscience. In support of "bottom-up theorizing", I explore the recent application of a new real-time molecular imaging technique (SCAPE microscopy) to mixture coding in olfaction. Seemingly emergent "higher-level" psychological effects in odor perception, irreducible to the physical stimulus, are linked back to underlying molecular mechanisms at the receptor level. The SCAPE study has notable theoretical impact. It provides a possible answer to the neurocomputational challenge in olfaction from combinatorial coding at the periphery: how does the brain discriminate different complex mixtures from widespread and overlapping receptor activation? The failure of previous reductionist structure-odor explanations is shown to reside in misconceptualizations of the critical causal elements involved. Causally fundamental features are not of parts independently of a mechanism. Components and their relevant features are units via their causal role within a mechanism. Here, new technologies allow revisiting our understanding of the ontology and levels of organization of a system.

Categorizing Smells: A Localist Approach

Humans are poorer at identifying smells and communicating about them, compared to other sensory domains. They also cannot easily organize odor sensations in a general conceptual space, where geometric distance could represent how similar or different all odors are. These two generalities are more or less accepted by psychologists, and they are often seen as connected: If there is no conceptual space for odors, then olfactory identification should indeed be poor. We propose here an important revision to this conclusion: We believe that the claim that there is no odor space is true only if by odor space, one means a conceptual space representing all possible odor sensations, in the paradigmatic sense used for instance for color. However, in a less paradigmatic sense, local conceptual spaces representing a given subset of odors do exist. Thus the absence of a global odor space does not warrant the conclusion that there is no olfactory conceptual map at all. Here we show how a localist account provides a new interpretation of experts and cross-cultural categorization studies: Rather than being exceptions to the poor olfactory identification and communication usually seen elsewhere, experts and cross-cultural categorization are here taken to corroborate the existence of local conceptual spaces.

Odor coding in the mammalian olfactory epithelium

Noses are extremely sophisticated chemical detectors allowing animals to use scents to interpret and navigate their environments. Odor detection starts with the activation of odorant receptors (ORs), expressed in mature olfactory sensory neurons (OSNs) populating the olfactory mucosa. Different odorants, or different concentrations of the same odorant, activate unique ensembles of ORs. This mechanism of combinatorial receptor coding provided a possible explanation as to why different odorants are perceived as having distinct odors. Aided by new technologies, several recent studies have found that antagonist interactions also play an important role in the formation of the combinatorial receptor code. These findings mark the start of a new era in the study of odorant-receptor interactions and add a new level of complexity to odor coding in mammals.

Metabolism of Odorant Molecules in Human Nasal/Oral Cavity Affects the Odorant Perception

In this study, we examined the mode of metabolism of food odorant molecules in the human nasal/oral cavity in vitro and in vivo. We selected 4 odorants, 2-furfurylthiol (2-FT), hexanal, benzyl acetate, and methyl raspberry ketone, which are potentially important for designing food flavors. In vitro metabolic assays of odorants with saliva/nasal mucus analyzed by gas chromatography mass spectrometry revealed that human saliva and nasal mucus exhibit the following 3 enzymatic activities: (i) methylation of 2-FT into furfuryl methylsulfide (FMS); (ii) reduction of hexanal into hexanol; and (iii) hydrolysis of benzyl acetate into benzyl alcohol. However, (iv) demethylation of methyl raspberry ketone was not observed. Real-time in vivo analysis using proton transfer reaction-mass spectrometry demonstrated that the application of 2-FT and hexanal through 3 different pathways via the nostril or through the mouth generated the metabolites FMS and hexanol within a few seconds. The concentration of FMS and hexanol in the exhaled air was above the perception threshold. A cross-adaptation study based on the activation pattern of human odorant receptors suggested that this metabolism affects odor perception. These results suggest that some odorants in food are metabolized in the human nasal mucus/saliva, and the resulting metabolites are perceived as part of the odor quality of the substrates. Our results help improve the understanding of the mechanism of food odor perception and may enable improved design and development of foods in relation to odor.

Development of a Piezoelectric-Based Odor Reproduction System

Odor reproduction, a branch of machine olfaction, is a technology through which a machine represents various odors by blending several odor sources in different proportions and releases them. In this paper, an odor reproduction system is proposed. The system includes an atomization-based odor dispenser using 16 micro-porous piezoelectric transducers. The authors propose the use of an electronic nose combined with a Principal Component Analysis–Linear Discriminant Analysis (PCA–LDA) model to evaluate the effectiveness of the system. The results indicate that the model can be used to evaluate the system.

A Critique of Olfactory Objects

Does the sense of smell involve the perception of odor objects? General discussion of perceptual objecthood centers on three criteria: stimulus representation, perceptual constancy, and figure-ground segregation. These criteria, derived from theories of vision, have been applied to olfaction in recent philosophical debates about psychology. An inherent problem with such framing of olfactory objecthood is that philosophers explicitly ignore the constitutive factors of the sensory systems that underpin the implementation of these criteria. The biological basis of odor coding is fundamentally different from the coding principles of the visual system. This article analyzes the three measures of perceptual objecthood against the biological background of the olfactory system. It contrasts the coding principles in olfaction with the visual system to show why these criteria of objecthood fail to be instantiated in odor perception. The argument demonstrates that olfaction affords perceptual categorization without the need to form odor objects.

Computational modeling of the olfactory receptor Olfr73 suggests a molecular basis for low potency of olfactory receptor-activating compounds

The mammalian olfactory system uses hundreds of specialized G-protein-coupled olfactory receptors (ORs) to discriminate a nearly unlimited number of odorants. Cognate agonists of most ORs have not yet been identified and potential non-olfactory processes mediated by ORs are unknown. Here, we used molecular modeling, fingerprint interaction analysis and molecular dynamics simulations to show that the binding pocket of the prototypical olfactory receptor Olfr73 is smaller, but more flexible, than binding pockets of typical non-olfactory G-protein-coupled receptors. We extended our modeling to virtual screening of a library of 1.6 million compounds against Olfr73. Our screen predicted 25 Olfr73 agonists beyond traditional odorants, of which 17 compounds, some with therapeutic potential, were validated in cell-based assays. Our modeling suggests a molecular basis for reduced interaction contacts between an odorant and its OR and thus the typical low potency of OR-activating compounds. These results provide a proof-of-principle for identifying novel therapeutic OR agonists.

Effects of point mutations in the binding pocket of the mouse major urinary protein MUP20 on ligand affinity and specificity

The mouse Major Urinary Proteins (MUPs) contain a conserved β-barrel structure with a characteristic central hydrophobic pocket that binds a variety of volatile compounds. After release of urine, these molecules are slowly emitted in the environment where they play an important role in chemical communication. MUPs are highly polymorphic and conformationally stable. They may be of interest in the construction of biosensor arrays capable of detection of a broad range of analytes. In this work, 14 critical amino acids in the binding pocket involved in ligand interactions were identified in MUP20 using in silico techniques and 7 MUP20 mutants were synthesised and characterised to produce a set of proteins with diverse ligand binding profiles to structurally different ligands. A single amino acid substitution in the binding pocket can dramatically change the MUPs binding affinity and ligand specificity. These results have great potential for the design of new biosensor and gas-sensor recognition elements.

How to be rational about empirical success in ongoing science: The case of the quantum nose and its critics

Empirical success is a central criterion for scientific decision-making. Yet its understanding in philosophical studies of science deserves renewed attention: Should philosophers think differently about the advancement of science when they deal with the uncertainty of outcome in ongoing research in comparison with historical episodes? This paper argues that normative appeals to empirical success in the evaluation of competing scientific explanations can result in unreliable conclusions, especially when we are looking at the changeability of direction in unsettled investigations. The challenges we encounter arise from the inherent dynamics of disciplinary and experimental objectives in research practice. In this paper we discuss how these dynamics inform the evaluation of empirical success by analyzing three of its requirements: data accommodation, instrumental reliability, and predictive power. We conclude that the assessment of empirical success in developing inquiry is set against the background of a model's interactive success and prospective value in an experimental context. Our argument is exemplified by the analysis of an apparent controversy surrounding the model of a quantum nose in research on olfaction. Notably, the public narrative of this controversy rests on a distorted perspective on measures of empirical success.

Functional odor classification through a medicinal chemistry approach

Crucial for any hypothesis about odor coding is the classification and prediction of sensory qualities in chemical compounds. The relationship between perceptual quality and molecular structure has occupied olfactory scientists throughout the 20th century, but details of the mechanism remain elusive. Odor molecules are typically organic compounds of low molecular weight that may be aliphatic or aromatic, may be saturated or unsaturated, and may have diverse functional polar groups. However, many molecules conforming to these characteristics are odorless. One approach recently used to solve this problem was to apply machine learning strategies to a large set of odors and human classifiers in an attempt to find common and unique chemical features that would predict a chemical's odor. We use an alternative method that relies more on the biological responses of olfactory sensory neurons and then applies the principles of medicinal chemistry, a technique widely used in drug discovery. We demonstrate the effectiveness of this strategy through a classification for esters, an important odorant for the creation of flavor in wine. Our findings indicate that computational approaches that do not account for biological responses will be plagued by both false positives and false negatives and fail to provide meaningful mechanistic data. However, the two approaches used in tandem could resolve many of the paradoxes in odor perception.

Recognition of the Component Odors in Mixtures

Natural olfactory stimuli are volatile-chemical mixtures in which relative perceptual saliencies determine which odor-components are identified. Odor identification also depends on rapid selective adaptation, as shown for 4 odor stimuli in an earlier experimental simulation of natural conditions. Adapt-test pairs of mixtures of water-soluble, distinct odor stimuli with chemical features in common were studied. Identification decreased for adapted components but increased for unadapted mixture-suppressed components, showing compound identities were retained, not degraded to individual molecular features. Four additional odor stimuli, 1 with 2 perceptible odor notes, and an added "water-adapted" control tested whether this finding would generalize to other 4-compound sets. Selective adaptation of mixtures of the compounds (odors): 3 mM benzaldehyde (cherry), 5 mM maltol (caramel), 1 mM guaiacol (smoke), and 4 mM methyl anthranilate (grape-smoke) again reciprocally unmasked odors of mixture-suppressed components in 2-, 3-, and 4-component mixtures with 2 exceptions. The cherry note of "benzaldehyde" (itself) and the shared note of "methyl anthranilate and guaiacol" (together) were more readily identified. The pervasive mixture-component dominance and dynamic perceptual salience may be mediated through peripheral adaptation and central mutual inhibition of neural responses. Originating in individual olfactory receptor variants, it limits odor identification and provides analytic properties for momentary recognition of a few remaining mixture-components.

Indistinguishable odour enantiomers: Differences between peripheral and central-nervous electrophysiological responses

The ability of humans to discriminate enantiomeric odour pairs is substance -specific. Current literature suggests that psychophysical discrimination of odour enantiomers mainly depends on the peripheral processing at the level of the olfactory sensory neurons (OSN). To study the influence of central processing in discrimination, we investigated differences in the electrophysiological responses to psychophysically indistinguishable (+)- and (-)- rose oxide enantiomers at peripheral and central-nervous levels in humans. We recorded the electro-olfactogram (EOG) from the olfactory epithelium and the EEG-derived olfactory event-related potentials (OERP). Results from a psychophysical three alternative forced choice test indicated indistinguishability of the two odour enantiomers. In a total of 19 young participants EOG could be recorded in 74 and OERP in 95% of subjects. Significantly different EOG amplitudes and latencies were recorded in response to the 2 stimuli. However, no such differences in amplitude or latency emerged for the OERP. In conclusion, although the pair of enantiomer could be discriminated at a peripheral level this did not lead to a central-nervous/cognitive differentiation of the two stimuli.

The Regular Interaction Pattern among Odorants of the Same Type and Its Application in Odor Intensity Assessment

The olfactory evaluation function (e.g., odor intensity rating) of e-nose is always one of the most challenging issues in researches about odor pollution monitoring. But odor is normally produced by a set of stimuli, and odor interactions among constituents significantly influenced their mixture’s odor intensity. This study investigated the odor interaction principle in odor mixtures of aldehydes and esters, respectively. Then, a modified vector model (MVM) was proposed and it successfully demonstrated the similarity of the odor interaction pattern among odorants of the same type. Based on the regular interaction pattern, unlike a determined empirical model only fit for a specific odor mixture in conventional approaches, the MVM distinctly simplified the odor intensity prediction of odor mixtures. Furthermore, the MVM also provided a way of directly converting constituents’ chemical concentrations to their mixture’s odor intensity. By combining the MVM with usual data-processing algorithm of e-nose, a new e-nose system was established for an odor intensity rating. Compared with instrumental analysis and human assessor, it exhibited accuracy well in both quantitative analysis (Pearson correlation coefficient was 0.999 for individual aldehydes (n = 12), 0.996 for their binary mixtures (n = 36) and 0.990 for their ternary mixtures (n = 60)) and odor intensity assessment (Pearson correlation coefficient was 0.980 for individual aldehydes (n = 15), 0.973 for their binary mixtures (n = 24), and 0.888 for their ternary mixtures (n = 25)). Thus, the observed regular interaction pattern is considered an important foundation for accelerating extensive application of olfactory evaluation in odor pollution monitoring.