Is It Possible to Predict the Odor of a Molecule on the Basis of its Structure?

Abstract

The olfactory sense is the dominant sensory perception for many animals. When Richard Axel and Linda B. Buck received the Nobel Prize in 2004 for discovering the G protein-coupled receptors' role in olfactory cells, they highlighted the importance of olfaction to the scientific community. Several theories have tried to explain how cells are able to distinguish such a wide variety of odorant molecules in a complex context in which enantiomers can result in completely different perceptions and structurally different molecules. Moreover, sex, age, cultural origin, and individual differences contribute to odor perception variations that complicate the picture. In this article, recent advances in olfaction theory are presented, and future trends in human olfaction such as structure-based odor prediction and artificial sniffing are discussed at the frontiers of chemistry, physiology, neurobiology, and machine learning.

Full text

International Journal of
Molecular Sciences
Review
Is It Possible to Predict the Odor of a Molecule on the
Basis of its Structure?
Manon Genva 1, Tierry Kenne Kemene 1, Magali Deleu 2, Laurence Lins 2
and Marie-Laure Fauconnier 1, *
1Laboratory of Chemistry of Natural Molecules, Gembloux Agro-Bio Tech, University of Liège,
5030 Gembloux, Belgium; m.genva@uliege.be (M.G.); kenne@gmx.com (T.K.K.)
2Laboratory of Molecular Biophysics at Interfaces, Gembloux Agro-Bio Tech, University of Liège,
5030 Gembloux, Belgium; Magali.Deleu@uliege.be (M.D.); L.Lins@uliege.be (L.L.)
*Correspondence: marie-laure.fauconnier@uliege.be; Tel.: +32-81-622-289
Received: 29 May 2019; Accepted: 18 June 2019; Published: 20 June 2019


Abstract:
The olfactory sense is the dominant sensory perception for many animals. When Richard
Axel and Linda B. Buck received the Nobel Prize in 2004 for discovering the G protein-coupled
receptors’ role in olfactory cells, they highlighted the importance of olfaction to the scientific
community. Several theories have tried to explain how cells are able to distinguish such a wide
variety of odorant molecules in a complex context in which enantiomers can result in completely
different perceptions and structurally different molecules. Moreover, sex, age, cultural origin, and
individual differences contribute to odor perception variations that complicate the picture. In this
article, recent advances in olfaction theory are presented, and future trends in human olfaction such
as structure-based odor prediction and artificial sniffing are discussed at the frontiers of chemistry,
physiology, neurobiology, and machine learning.
Keywords:
olfaction; odor; sensory perception; flavor; volatile organic compounds; structure-function
relationship; olfactory sense; odorant
1. Introduction
1.1. The Crucial Roles of the Olfactory Sense
It seems rational that the ability to recognize odor is common to every animal that has a nose,
but this may become less obvious in single-cell micro-organisms. However, almost all organisms,
including micro-organisms, can sense environmental cues through chemoreception [
1
]. The ability to
perceive volatile organic compounds (VOCs) is found in various phyla from single-cell organisms to
more complex ones such as mammals [
2
]. For most animals, the sense of smell is vital; the roles of odor
perception can be related to various functions involved in guaranteeing survival and reproduction,
such as finding food, avoiding danger, conspecific recognition, and searching for a mate [
3
–
6
]. Insects
and certain vertebrate animals locate their prey through smell. To illustrate this, mosquitos locate
mammals in order to suck their blood by detecting carbon dioxide, while common predators, especially
felines, can recognize the scent of blood hundreds of miles away [
7
]. The sense of smell also allows
animals to avoid danger by detecting foul-smelling food that may contain poisons and pathogens [
8
].
Finally, smell helps to build strong and positive ties between newborn animals, including humans, and
their mothers. During pregnancy, women develop a particular pattern of volatile molecules which are
useful to newborns to recognize their mother and to guide feeding immediately after birth and the first
weeks of life [9].
Int. J. Mol. Sci. 2019,20, 3018; doi:10.3390/ijms20123018 www.mdpi.com/journal/ijms

Int. J. Mol. Sci. 2019,20, 3018 2 of 16
1.2. Human Response to Odor
Smell is not simply a biological and psychological experience; it is also a social and cultural
phenomenon [
10
], as the sense of smell differs depending on culture, age, gender, and health status [
11
].
Olfactory perception is widely influenced by background and semantic information, as two people
with different cultural backgrounds can have different reactions when smelling the same thing. For
example, Canadians can better describe the scent of maple than French people, the latter being better
at describing lavender [
12
]. Odor recognition is also influenced by age, as children (<16 years old) and
older people (>55 years old) are less sensitive to odor than young and middle-aged adults [
13
–
16
].
Variation in olfactory perception depending on sex is less obvious [
13
]. In a recent meta-analysis, it
was shown that women generally outperform men in the identification and discrimination of odorants.
Ladies usually also have lower threshold values for odorant molecules. Nevertheless, even if the
differences are statistically significant, they are small [17].
1.3. Physical and Chemical Features of Odor
According to Richard Axel and Linda B. Buck (recipients of the 2004 Nobel Prize in medicine or
physiology), smell is a chemical stimulus with a physiological response which can be due to one or a
combination of odorant molecules [
18
]. The latter are volatilized chemical compounds, usually at small
concentrations, that are perceived by animals, including humans, through the sense of olfaction [
19
].
The term “odor” is frequently used to name a scent that may or may not be pleasant, while the terms
“fragrance” and “aroma” are mostly used by the cosmetic and food industries to describe a pleasant
odor [19].
Odorant molecules include VOCs, which constitute a large class of low-molecular-weight (<300 Da)
carbon-containing compounds characterized by their high vapor pressure (
≥
0.01 kPa at 20
◦
C) and
high-to-moderate hydrophobicity [
20
]. An odor primarily originates from a compound volatilizing at
ambient temperature and thus reaching the nose. However, the vapor pressure value is not sufficient
to predict whether a compound is odorant or not. For example, humans, on the one hand, can
smell NO
2
but not CO
2
; on the other hand, they can smell relatively large molecules such as musk
compounds [
21
,
22
]. Odorant molecules are not limited to carbon-containing compounds, as both
organic and inorganic molecules may have a smell. For example, ammonia (NH
3
) is an inorganic
compound that has a distinctive fishy scent [
23
]. Elemental chlorine gas (Cl
2
) has an acrid smell.
Hydrogen sulfide (H
2
S) is another inorganic odorant with a rotten egg scent. It is widely accepted
that humans are able to detect the presence of functional groups more easily with great reliability
than a single elemental compound. This is the case of thiols (-SH), oxines (-NOH) and nitro groups
(-NO
2
), respectively known for their sulfurous odor, green camphoraceous smell, and sweet ethereal
character [23].
1.4. Odor and Structure Relationship
Odorant compounds that have the same functional group seem to have similar odors, as esters
have a fruity and floral smell, lactones display a coconut or apricot character, amines have an
animal/roasted scent, thiols have a rotten or alliaceous smell [
24
], volatile fatty acids have a sour to
rancid smell, and aldehydes are associated with green odors such as grass cuttings or leaves [
25
,
26
].
Many studies have focused on the relationship between the molecular structure of a compound and its
odor [
27
–
29
]. Indeed, resolving this issue could have a significant economic impact on perfume and
aroma formulations. Some functional groups and their related smells are presented in Figure 1.

Int. J. Mol. Sci. 2019,20, 3018 3 of 16
Int.J.Mol.Sci.2019,20,xFORPEERREVIEW3of17
Int.J.Mol.Sci.2019,20,x;doi:FORPEERREVIEWwww.mdpi.com/journal/ijms
Functional
relatedgroupOdor‐typeandrelatedcompound
Citrusy
Aldehydes
CH
3
O

Sweet,aldehydic,waxy,
orangepeel,citrus,floral
decanal
CH3
O
CH3

Fresh,amber,aldehydic,
moss,citrus,tuberose,
metallic,waxy,coumarinic
2‐methylundecanal
CH
3
O
CH
3

Fresh,dry,citrus,waxy,
watery
2‐methyldecanal
Fishy
AminesCH3
NH2

Fishy,amine‐like
heptan‐1‐amine
CH
3
NH
CH
3

Ammoniacal,fishy,musty
N‐butylbutan‐1‐amine
Fruity
Esters
OO
CH
3
O
O
CH
3

Sweet,fruity,apple,green,
tropical,plum,woody
ethyl(2‐methyl‐1,3‐dioxolan‐2‐yl)acetate
CH
3
CH
3
CH
3
O
O
CH
3

Fruity,woody,green,apple,
herbal
2‐tert‐butylcyclohexylacetate
CH
3
O
O
CH
3

Sweet,fruity,tutti‐frutti,
lifting,diffusive,apple
ethylbutanoate
Figure1.Examplesofodorantswithcommonfunctionalgroupsandsimilarodors[23].
Figure 1. Examples of odorants with common functional groups and similar odors [23].
However, only a few odors can be predicted based on the functional group of the molecule.
Indeed, some odorant molecules have the same functional group but different odors. This
is the case, for example, for 4,4-dimethyl-2-octeno-
δ
-lactone, 8-methyl-2-noneno-
δ
-lactone, and
5,6,6-trimethyl-2-hepteno-
δ
-lactone (Figure 2), which are three lactones that have very close structures
but very distinctive smells. In fact, the first molecule has a minty odor, the second a buttery odor, and
the last a terpene-like and camphorous character [30].
Likewise, enantiomeric compounds, also known as optical isomers, obviously have the same
chemical functions and are structurally close, but only as few as 5% of enantiomer couples have
a similar smell [
31
]. A very common example used by many organic chemistry teachers is the
two enantiomers of limonene: (S)-(-)-limonene smells like lemons, while (R)-(+)-limonene has the

Int. J. Mol. Sci. 2019,20, 3018 4 of 16
characteristic smell of orange. Another well-known example of an enantiomer couple with nonidentical
smells is (R)-γ-methylcyclogeranate, which smells like camphor, and (S)-γ-methyl cyclogeranate, the
scent of which is described as fruity [
31
]. When the molecule has two chiral centers, four different
enantiomers exist, potentially leading to four different odors (e.g., mentha-8-thiol-3-ones) (Figure 3) [
32
].
Int.J.Mol.Sci.2019,20,xFORPEERREVIEW4of17
However,onlyafewodorscanbepredictedbasedonthefunctionalgroupofthemolecule.
Indeed,someodorantmoleculeshavethesamefunctionalgroupbutdifferentodors.Thisisthecase,
forexample,for4,4‐dimethyl‐2‐octeno‐δ‐lactone,8‐methyl‐2‐noneno‐δ‐lactone,and
5,6,6‐trimethyl‐2‐hepteno‐δ‐lactone(Figure2),whicharethreelactonesthathaveveryclose
structuresbutverydistinctivesmells.Infact,thefirstmoleculehasamintyodor,theseconda
butteryodor,andthelastaterpene‐likeandcamphorouscharacter[30].
4,4‐dimethyl‐2‐octeno‐δ‐lactone
OO
CH
3
CH
3
CH
3

Minty
8‐methyl‐2‐noneno‐δ‐lactone
OO
CH
3
CH
3

Buttery
5,6,6‐trimethyl‐2‐hepteno‐δ‐lactone
OO
CH
3
CH
3
CH
3
CH
3

Terpene‐like,
campherous
Figure2.Examplesofodorantswithcommonfunctionalgroupsanddissimilarodors[24].
Likewise,enantiomericcompounds,alsoknownasopticalisomers,obviouslyhavethesame
chemicalfunctionsandarestructurallyclose,butonlyasfewas5%ofenantiomercoupleshavea
similarsmell[31].Averycommonexampleusedbymanyorganicchemistryteachersisthetwo
enantiomersoflimonene:(S)‐(‐)‐limonenesmellslikelemons,while(R)‐(+)‐limonenehasthe
characteristicsmelloforange.Anotherwell‐knownexampleofanenantiomercouplewith
nonidenticalsmellsis(R)‐γ‐methylcyclogeranate,whichsmellslikecamphor,and(S)‐γ‐methyl
cyclogeranate,thescentofwhichisdescribedasfruity[31].Whenthemoleculehastwochiral
centers,fourdifferentenantiomersexist,potentiallyleadingtofourdifferentodors(e.g.,
mentha‐8‐thiol‐3‐ones)(Figure3)[32].
Figure 2. Examples of odorants with common functional groups and dissimilar odors [24].
Int.J.Mol.Sci.2019,20,xFORPEERREVIEW5of17
Int.J.Mol.Sci.2019,20,x;doi:FORPEERREVIEWwww.mdpi.com/journal/ijms
CH
3
CH
3
CH
2

CH
3
CH
3
CH
2

(S)‐(‐)‐limonene(R)‐(+)‐limonene
LemonOrange
CH
3
CH
3
CH
3
O
O
CH
3

CH
3
CH
3
CH
3
O
O
CH
3

(R)‐γ‐methylcyclogeranate(S)‐γ‐methylcyclogeranate
CamphoraceousFruity
CH
3
OSH
CH
3
CH
3

CH
3
OSH
CH
3
CH
3

(1S,4R)‐cis‐mentha‐8‐thiol‐3‐one(1R,4S)‐cis‐mentha‐8‐thiol‐3‐one
Tropical,fruityRuber
CH
3
OSH
CH
3
CH
3

CH
3
OSH
CH
3
CH
3

(1R,4R)‐trans‐mentha‐8‐thiol‐3‐one(1S,4S)‐trans‐mentha‐8‐thiol‐3‐one
Onion‐likeTropical,sulfurous
Figure3.Examplesofenantiomericcompoundswithdissimilarodors[31,33].
Finally,thereisthecaseofstructurallydifferentorganiccompoundshavingasimilarsmell;for
example,musk‐relatedodors(Figure4)[32,34–37].
Figure 3. Examples of enantiomeric compounds with dissimilar odors [31,33].

Int. J. Mol. Sci. 2019,20, 3018 5 of 16
Finally, there is the case of structurally different organic compounds having a similar smell; for
example, musk-related odors (Figure 4) [32,34–37].
Int.J.Mol.Sci.2019,20,xFORPEERREVIEW6of17
Int.J.Mol.Sci.2019,20,x;doi:FORPEERREVIEWwww.mdpi.com/journal/ijms
O
CH
3

CH
3
O
CH
3
CH
3
NO
2
CH
3
CH
3
CH
3
O
2
N

3‐methylcyclopentadecanone1‐(4‐tert‐butyl‐2,6‐dimethyl‐3,5‐dinitrophe
nyl)ethanone
CH
3
O
CH
3
CH
3
CH
3
CH
3
CH
3
CH
3

CH3
CH3
CH3
O
CH3
CH3
O
O
CH3


1‐[1,1,2,6‐tetramethyl‐3‐(propan‐2‐yl)‐2,3‐dihydroinde
n‐5‐yl]ethanone
2‐(1‐(3,3‐dimethylcyclohexyl)ethoxy)‐2‐me
thylpropylpropionate
Figure4.Examplesofodorantswithdifferentstructuresandsimilarodors(musk)[32,37].
Besideschangingtheodor,asmallstructuralchangetoamoleculemaycauseadecreasein
odorintensity.Asanexample(Figure5),theleftmoleculehasaurinoussmell,whilethemolecule
ontheright,whichonlydisplaysoneextraCH3,isodorless[37,38].
CH
3
O
CH
3
CH
3
CH
3
CH
3

4‐(trans‐4‐tert‐butylcyclohexyl)pentan‐2‐one
Urinous
CH
3
O
CH
3
CH
3
CH
3
CH
3
CH
3

4‐(trans‐4‐tert‐butylcyclohexyl)‐4‐methylpentan‐2‐one
Odorless
Figure5.Exampleofthechangingodorantcharacterofacompoundwithslightstructural
modification.
Asmallchangeinthestructureorfunctionalgroupofamoleculecansignificantlyalteritssmell
inamannerthatacurrentpredictionodor–structuremodelcannotcompletelyexplain.
2.OlfactionMechanisms
2.1.ThePhysiologyofOlfaction
Theolfactorysystem’sorganizationisremarkablysimilarinvariousanimals,frominsectsto
mammals,allowingthedetectionofalargearrayofstructurallydifferentmolecules.Themechanism
ofolfactioncanbedividedintofourmainsteps:airflowoftheodorants,bindingtoreceptors,
Figure 4. Examples of odorants with different structures and similar odors (musk) [32,37].
Besides changing the odor, a small structural change to a molecule may cause a decrease in odor
intensity. As an example (Figure 5), the left molecule has a urinous smell, while the molecule on the
right, which only displays one extra CH3, is odorless [37,38].
Int.J.Mol.Sci.2019,20,xFORPEERREVIEW6of17
Int.J.Mol.Sci.2019,20,x;doi:FORPEERREVIEWwww.mdpi.com/journal/ijms
O
CH
3

CH
3
O
CH
3
CH
3
NO
2
CH
3
CH
3
CH
3
O
2
N

3‐methylcyclopentadecanone1‐(4‐tert‐butyl‐2,6‐dimethyl‐3,5‐dinitrophe
nyl)ethanone
CH
3
O
CH
3
CH
3
CH
3
CH
3
CH
3
CH
3

CH3
CH3
CH3
O
CH3
CH3
O
O
CH3


1‐[1,1,2,6‐tetramethyl‐3‐(propan‐2‐yl)‐2,3‐dihydroinde
n‐5‐yl]ethanone
2‐(1‐(3,3‐dimethylcyclohexyl)ethoxy)‐2‐me
thylpropylpropionate
Figure4.Examplesofodorantswithdifferentstructuresandsimilarodors(musk)[32,37].
Besideschangingtheodor,asmallstructuralchangetoamoleculemaycauseadecreasein
odorintensity.Asanexample(Figure5),theleftmoleculehasaurinoussmell,whilethemolecule
ontheright,whichonlydisplaysoneextraCH3,isodorless[37,38].
CH
3
O
CH
3
CH
3
CH
3
CH
3

4‐(trans‐4‐tert‐butylcyclohexyl)pentan‐2‐one
Urinous
CH
3
O
CH
3
CH
3
CH
3
CH
3
CH
3

4‐(trans‐4‐tert‐butylcyclohexyl)‐4‐methylpentan‐2‐one
Odorless
Figure5.Exampleofthechangingodorantcharacterofacompoundwithslightstructural
modification.
Asmallchangeinthestructureorfunctionalgroupofamoleculecansignificantlyalteritssmell
inamannerthatacurrentpredictionodor–structuremodelcannotcompletelyexplain.
2.OlfactionMechanisms
2.1.ThePhysiologyofOlfaction
Theolfactorysystem’sorganizationisremarkablysimilarinvariousanimals,frominsectsto
mammals,allowingthedetectionofalargearrayofstructurallydifferentmolecules.Themechanism
ofolfactioncanbedividedintofourmainsteps:airflowoftheodorants,bindingtoreceptors,
Figure 5.
Example of the changing odorant character of a compound with slight structural modification.
A small change in the structure or functional group of a molecule can significantly alter its smell
in a manner that a current prediction odor–structure model cannot completely explain.
2. Olfaction Mechanisms
2.1. The Physiology of Olfaction
The olfactory system’s organization is remarkably similar in various animals, from insects to
mammals, allowing the detection of a large array of structurally different molecules. The mechanism of
olfaction can be divided into four main steps: airflow of the odorants, binding to receptors, transduction
of the signal, and information processing [39]. Hereafter, we focus on human olfaction (Figure 6).
Odor molecules can reach the nasal cavity either through orthonasal olfaction (direct inhalation
in front of the nose) or through the throat when the tongue pushes air to the back of the nasal cavity
while chewing or drinking (retronasal olfaction) [33].
Inside the nasal cavity, odor perception is due to the interaction of volatile compounds with the
olfactory receptor neurons (ORNs) that lie in the olfactory epithelium, which occupies a 3.7 cm
2
zone

Int. J. Mol. Sci. 2019,20, 3018 6 of 16
in the upper part of the nasal cavity. The epithelium is covered by mucus and contains olfactory glands
that secrete the enzymes found in the mucus [
40
]. Humans have around 12 million ORNs in each
epithelium (right and left); as the olfaction system is bilateral, there are two of each structure [41].
Int.J.Mol.Sci.2019,20,xFORPEERREVIEW7of17
transductionofthesignal,andinformationprocessing[39].Hereafter,wefocusonhumanolfaction
(Figure6).
Odormoleculescanreachthenasalcavityeitherthroughorthonasalolfaction(directinhalation
infrontofthenose)orthroughthethroatwhenthetonguepushesairtothebackofthenasalcavity
whilechewingordrinking(retronasalolfaction)[33].
Insidethenasalcavity,odorperceptionisduetotheinteractionofvolatilecompoundswiththe
olfactoryreceptorneurons(ORNs)thatlieintheolfactoryepithelium,whichoccupiesa3.7cm
2
zone
inthe

upperpartofthenasalcavity.Theepitheliumiscoveredbymucusandcontainsolfactory
glandsthatsecretetheenzymesfoundinthemucus[40].Humanshavearound12millionORNsin
eachepithelium(rightandleft);astheolfactionsystemisbilateral,therearetwoofeachstructure
[41].
Sensorytransductionisthemechanismbywhichtheinformationrelatedtotheodordetected
byanORNistransmittedtothebrain.Infact,eachORNhasatoneend20–30ciliabathinginmucus
andcontainingolfactoryreceptors(ORs)thatbindtoodormoleculestogiveanelectricalresponse.
TheelectricalmessageistransferredthroughtheaxonsituatedattheotherendoftheORNstonerve
fiberssituatedatthebackofthenasalcavity[33,40].EachORNexpressesonereceptortypein
humans,whereasitwasshownthatasingleneuronmayexpresstwodifferentreceptortypesin
otherspecies[41].TheaxonsofORNspassthroughtheethmoidbonetoformglomeruli.Glomeruli
arecomposedofORNaxonterminalsandthedendritesofmitralcells(second‐orderneurons).The
axonsofalltheORNscontainingthesametypeofodorreceptoraregroupedinthesame
glomerulus.Inhumans,morethan5500glomeruliarepresent,formingtheolfactorybulb.The
olfactorybulbistheplacewherethesignalistreatedandtransmittedtothebrain[41,42].Axonsof
themitralcellsaregroupedinolfactorytractsconnectedtotheolfactorycortex.Untilrecently,the
olfactorysystemwasthoughttobeipsilateral:theleftepitheliumisconnectedtotheleftbulbwhich
itselfisconnectedtotheleftcortex,butitisnowknownthattheconnectivitycanalsobe
contralateral[41].Followingolfactorystimuli,ithasbeenshownthatseveralbrainzonesare
activated,suchasthethalamus,amygdala,andorbitofrontalcortex,makingolfactionacomplex
sensoryexperience[33].

Figure6.Mechanismsofhumanolfaction.1.Orthonasalolfaction;2.retronasalolfaction;3.nasal
cavity;4.olfactorybulb;A.mitralcell;B.glomerulus;C.axon;D.cilia;E.olfactoryreceptor;F.
olfactoryreceptorneuron.
2.2.OlfactionTheory:StudyingStructure–OdorRelationships
Therelationshipbetweenthemolecularstructureofacompoundanditsodorhasbeenthe
subjectofmanystudies.Inthisrespect,twomainschoolsofthoughtcoexist:theglobalapproach
andthestudyofodorantmolecule–receptorinteraction.Theglobalapproachisquitedescriptive
Figure 6.
Mechanisms of human olfaction.
1.
Orthonasal olfaction;
2.
retronasal olfaction;
3.
nasal
cavity; 4. olfactory bulb;
A.
mitral cell;
B.
glomerulus; C. axon;
D.
cilia; E. olfactory receptor;
F.
olfactory
receptor neuron.
Sensory transduction is the mechanism by which the information related to the odor detected by
an ORN is transmitted to the brain. In fact, each ORN has at one end 20–30 cilia bathing in mucus
and containing olfactory receptors (ORs) that bind to odor molecules to give an electrical response.
The electrical message is transferred through the axon situated at the other end of the ORNs to nerve
fibers situated at the back of the nasal cavity [
33
,
40
]. Each ORN expresses one receptor type in humans,
whereas it was shown that a single neuron may express two different receptor types in other species [
41
].
The axons of ORNs pass through the ethmoid bone to form glomeruli. Glomeruli are composed of
ORN axon terminals and the dendrites of mitral cells (second-order neurons). The axons of all the
ORNs containing the same type of odor receptor are grouped in the same glomerulus. In humans,
more than 5500 glomeruli are present, forming the olfactory bulb. The olfactory bulb is the place
where the signal is treated and transmitted to the brain [
41
,
42
]. Axons of the mitral cells are grouped
in olfactory tracts connected to the olfactory cortex. Until recently, the olfactory system was thought
to be ipsilateral: the left epithelium is connected to the left bulb which itself is connected to the left
cortex, but it is now known that the connectivity can also be contralateral [
41
]. Following olfactory
stimuli, it has been shown that several brain zones are activated, such as the thalamus, amygdala, and
orbitofrontal cortex, making olfaction a complex sensory experience [33].
2.2. Olfaction Theory: Studying Structure–Odor Relationships
The relationship between the molecular structure of a compound and its odor has been the
subject of many studies. In this respect, two main schools of thought coexist: the global approach
and the study of odorant molecule–receptor interaction. The global approach is quite descriptive
and is ineffective at explaining the mechanisms underlying the olfaction chain. Nevertheless, these
studies can be helpful for flavorists and perfumers in search of new raw materials for the products they
develop [
32
,
34
–
36
,
43
–
46
]. The complete olfactory chain is complex and can lead to several biases. First
of all, the way sniffing is performed, either rapidly or slowly, will influence the odor concentration
reaching the nostril, but it seems that the olfactory bulb is able to distinguish odor concentration
variation and nasal flow modification (mechanical perception) [
41
,
47
]. Retronasal and orthonasal
olfaction can result in different perceptions, mainly due to the differential solubility of the odorants in
the nasal or nasopharyngeal mucus, which differ in composition [
48
]. Prereceptor events, such as the
enzymatic conversion of odorants in nasal mucus and binding to odorant-binding proteins, will also
affect odor perception (e.g., conversion of aldehydes and esters to acids and alcohols) [
49
,
50
]. The nasal
mucus is rich in odorant-binding proteins that have variable affinity for odorants depending on their

Int. J. Mol. Sci. 2019,20, 3018 7 of 16
chemical structure. The precise role of these odorant-binding proteins is quite hypothetical: trapping
of highly concentrated odorants to protect the nasal epithelium, transporting hydrophobic compounds
through the mucus, and removing odorants from the receptors to allow fast reactivation [
51
,
52
].
Post-receptor events can also complicate the understanding of the relationship between chemical
structure and odor perception. Indeed, the transformation by the brain of the signal originating from
the nose is not completely understood. The glomeruli are described as a temporally well-organized
unit playing an important role in signal transformation that can be compared to the thalamus’ role in
other sensory pathways; for example, helping to discriminate odor concentration changes for odor
source localization [
53
–
56
]. Another problem emerges once the signal has been treated by the brain;
that is, expressing the perceived odor in a way that everyone can understand. Unlike sight and hearing,
which can be reduced to easily measurable physical parameters, olfaction is subjective, and personal
characterization renders information sharing difficult even among professionals. Different reference
scales such as the “field of odors” have been created using reference molecules to classify odors but are
still far from a universal language for odor description [57,58].
Studies based on odorant–receptor interactions allow for a subtler comprehension of olfaction
mechanisms but require extremely pure compounds, since a trace of a minor enantiomer can impair
the quality of the conclusions. One must keep in mind that some odorants have a very low perception
threshold; for example, the ethyl mercaptan added to propane as a warning agent can be detected at a
concentration of 10
−7
ng/L, the lowest one being pyrrolidino[1,2-e]-4H-2,4-dimethyl-1,3,5-dithiazine,
the typical flavor of cooked shellfish, at 10−8ng/L for the two isomers [32,41].
2.3. Odorant–Receptor Interactions
In 2004, Richard Axel and Linda B. Buck were awarded the Nobel Prize in medicine or physiology
for their research on odorant receptors and the organization of the olfactory system [
18
]. These studies
allowed a better understanding of the human smell mechanism through an elucidation of the olfactory
system [
18
]. Before that, the sense of smell was less well understood among our senses, even though
numerous different theories had been proposed [
33
]. Without going into deeper explanation, almost
30 theories [
59
,
60
] had been proposed regarding the olfactory sense before Richard Axel and Linda B.
Buck’s discovery, including a model for the stimulation of the organ of smell based on the polarization
effect [
61
], a model for the olfactory membrane [
62
], the piezoelectric effect [
63
], olfactory transduction
based on the formation of a weak bound complex between an absorbed gas molecule and certain
carotenoid pigments [
64
], an analogy with polar compounds in chromatographic systems [
65
], and a
mechanism based on inelastic electron spectroscopy [
59
]. Among these theories, the stereochemical
theory of olfaction, known as Amoore’s theory, had been very popular for quite some time [
66
].
Based on seven primary odors (e.g., camphor and flora), a general shape was associated with each
primary odor. As an example, a camphoraceous odor was linked to a hydrophobic, ellipsoidal shape
with a long axis of 0.95 nm and a short axis of 0.75 nm. This theory allowed the design of numerous
commercially successful novel molecules such as Javanol
®
, Rossitol
®
, Belambre
®
, and Azurone
®
, but
when the number of described specific anosmia (i.e., the inability to perceive specific odors) increased
from 7, which was coherent with the model, to 90, the theory was discarded [
66
–
68
]. On the other hand,
the vibrational theory considers that the detection of odorants is due to their vibrational frequencies
and not to their shape; an electron transfer can occur between odorants and the active site of ORs.
Several studies showing that humans are able to differentiate isotopomers (isomers with isotopic atoms)
reinforced the theory because the latter differ in their molecular vibrational frequencies. However,
in a recent study, Block et al. tested series of isotopomers of musk and other compounds for their
capacity to activate specific odorant receptors
in vitro
and found no experimental evidence supporting
the vibrational theory. Peri-receptor events or odorant contaminations, rather than vibrational effects
at the receptor level, are suggested to explain the ability to discriminate isotopomers. It has since been
proposed that the study of the nature of odorant–receptor recognition should be based on receptor
activation mechanisms rather than odor perception [59].

Int. J. Mol. Sci. 2019,20, 3018 8 of 16
Buck and Axel hypothesized that the binding of odor molecules to specific surface receptors
(GTP-binding protein-coupled receptors, or GPCR) activates specific G proteins. GPCRs are
transmembrane proteins that cross the membrane seven times with an extracellular N-terminus
and an intracellular C-terminus (seven
α
-helices and six loops) [
18
,
33
]. Because the direct identification
of those transmembrane proteins is difficult in biological samples, 18 different members of a large
multigene family only expressed in the olfactory epithelium were cloned and characterized [
18
].
Genetic studies have shown that OR proteins are mainly expressed in ORNs and are highly variable,
especially in the transmembrane domains—3, 4, and 5, presumably—where the odorant binds the
protein [
18
]. Figure 7presents the cascade of reactions occurring when an odorant molecule enters the
nasal cavity. The odorant crosses the mucus directly or via transport proteins and then reaches the
odor receptor. After the binding of the odorant, the receptor undergoes structural modifications and
activates G proteins that are located on the cytoplasmic side. An active subunit called Ga is liberated,
activating, in turn, the lyase adenylate cyclase, resulting in the conversion of adenosine triphosphate
(ATP) into cyclic adenosine monophosphate (cAMP). The cAMP is able to open cyclic nucleotide-gated
ion channels allowing Ca
2+
and Na
+
ions to enter the cell, which results in the depolarization of the
ORN and transmission of the information to the brain. Nevertheless, the entry of Ca
++
ions opens
Ca++-dependent Cl-channels, which increases the depolarization [18,33,69,70].
Int.J.Mol.Sci.2019,20,xFORPEERREVIEW9of17
describedspecificanosmia(i.e.,theinabilitytoperceivespecificodors)increasedfrom7,whichwas
coherentwiththemodel,to90,thetheorywasdiscarded[66–68].Ontheotherhand,thevibrational
theoryconsidersthatthedetectionofodorantsisduetotheirvibrationalfrequenciesandnottotheir
shape;anelectrontransfercanoccurbetweenodorantsandtheactivesiteofORs.Severalstudies
showingthathumansareabletodifferentiateisotopomers(isomerswithisotopicatoms)reinforced
thetheorybecausethelatterdifferintheirmolecularvibrationalfrequencies.However,inarecent
study,Blocketal.testedseriesofisotopomersofmuskandothercompoundsfortheircapacityto
activatespecificodorantreceptorsinvitroandfoundnoexperimentalevidencesupportingthe
vibrationaltheory.Peri‐receptoreventsorodorantcontaminations,ratherthanvibrationaleffectsat
thereceptorlevel,aresuggestedtoexplaintheabilitytodiscriminateisotopomers.Ithassincebeen
proposedthatthestudyofthenatureofodorant–receptorrecognitionshouldbebasedonreceptor
activationmechanismsratherthanodorperception[59].
BuckandAxelhypothesizedthatthebindingofodormoleculestospecificsurfacereceptors
(GTP‐bindingprotein‐coupledreceptors,orGPCR)activatesspecificGproteins.GPCRsare
transmembraneproteinsthatcrossthemembraneseventimeswithanextracellularN‐terminusand
anintracellularC‐terminus(sevenα‐helicesandsixloops)[18,33].Becausethedirectidentification
ofthosetransmembraneproteinsisdifficultinbiologicalsamples,18differentmembersofalarge
multigenefamilyonlyexpressedintheolfactoryepitheliumwereclonedandcharacterized[18].
GeneticstudieshaveshownthatORproteinsaremainlyexpressedinORNsandarehighlyvariable,
especiallyinthetransmembranedomains—3,4,and5,presumably—wheretheodorantbindsthe
protein[18].Figure7presentsthecascadeofreactionsoccurringwhenanodorantmoleculeenters
thenasalcavity.Theodorantcrossesthemucusdirectlyorviatransportproteinsandthenreaches
theodorreceptor.Afterthebindingoftheodorant,thereceptorundergoesstructuralmodifications
andactivatesGproteinsthatarelocatedonthecytoplasmicside.AnactivesubunitcalledGais
liberated,activating,inturn,thelyaseadenylatecyclase,resultingintheconversionofadenosine
triphosphate(ATP)intocyclicadenosinemonophosphate(cAMP).ThecAMPisabletoopencyclic
nucleotide‐gatedionchannelsallowingCa
2+
andNa
+
ionstoenterthecell,whichresultsinthe
depolarizationoftheORNandtransmissionoftheinformationtothebrain.Nevertheless,theentry
ofCa
++
ionsopensCa
++
‐dependentCl
‐
channels,whichincreasesthedepolarization[18,33,69,70].

Figure7.Cascadeofreactionsoccurringwhenanodorantmoleculeentersthenasalcavity.Odorants
crossthemucusdirectlyorviatransportproteins(Tp)andthenreachtheodorreceptor,which
Figure 7.
Cascade of reactions occurring when an odorant molecule enters the nasal cavity. Odorants
cross the mucus directly or via transport proteins (Tp) and then reach the odor receptor, which
undergoes structural modifications and activates G proteins (G). An active subunit (Ga) is liberated,
activating the lyase adenylate cyclase, resulting in the conversion of adenosine triphosphate (ATP)
into cyclic adenosine monophosphate (cAMP). The cAMP is able to open cyclic nucleotide-gated ion
channels, allowing Ca
2+
and Na
+
ions to enter the cell, which results in the depolarization of the odor
receptor neuron (ORN) and transmission of the information to the brain.
The number of odorant molecules that humans are able to distinguish is huge, but their exact
number is still controversial, with estimates ranging from 400,000 to 1 million [
2
]. Two theories could
explain this ability: a large number of olfactory receptors, each interacting with a few odorants, or
a small number of olfactory receptors responding to a great number of odorants in a combinatorial
way. With only 396 unique olfactory receptors [
71
], only the second hypothesis is plausible, which is
considerably different than the classical “lock and key” mechanism [
2
]. Compared with other mammals
such as rodents, humans have fewer olfactory receptors (1130 in mice) [
71
] and smaller olfactory bulbs,
but the number of neurons they contain is comparable [72].
Rather than binding specific ligands, olfactory receptors have affinity for a range of odorants, and
an odorant molecule can bind several receptors with varying affinities depending on its physicochemical

Int. J. Mol. Sci. 2019,20, 3018 9 of 16
properties, such as the molecular volume. Likewise, the identification of a particular odorant is not
due to the activation of one single receptor but to the activation of a group of receptors with a special
pattern which is typical for combinatorial coding. It has been shown that receptors can be broadly
tuned and respond to many different odorants, being most responsive to structurally similar odorants,
or narrowly tuned, responding to a small group of odorants, as depicted in Figure 8[23].
Int.J.Mol.Sci.2019,20,xFORPEERREVIEW10of17
undergoesstructuralmodificationsandactivatesGproteins(G).Anactivesubunit(Ga)isliberated,
activatingthelyaseadenylatecyclase,resultingintheconversionofadenosinetriphosphate(ATP)
intocyclicadenosinemonophosphate(cAMP).ThecAMPisabletoopencyclicnucleotide‐gatedion
channels,allowingCa
2+
andNa
+
ionstoenterthecell,whichresultsinthedepolarizationoftheodor
receptorneuron(ORN)andtransmissionoftheinformationtothebrain.
Thenumberofodorantmoleculesthathumansareabletodistinguishishuge,buttheirexact
numberisstillcontroversial,withestimatesrangingfrom400,000to1million[2].Twotheoriescould
explainthisability:alargenumberofolfactoryreceptors,eachinteractingwithafewodorants,ora
smallnumberofolfactoryreceptorsrespondingtoagreatnumberofodorantsinacombinatorial
way.Withonly396uniqueolfactoryreceptors[71],onlythesecondhypothesisisplausible,whichis
considerablydifferentthantheclassical“lockandkey”mechanism[2].Comparedwithother
mammalssuchasrodents,humanshavefewerolfactoryreceptors(1130inmice)[71]andsmaller
olfactorybulbs,butthenumberofneuronstheycontainiscomparable[72].
Ratherthanbindingspecificligands,olfactoryreceptorshaveaffinityforarangeofodorants,
andanodorantmoleculecanbindseveralreceptorswithvaryingaffinitiesdependingonits
physicochemicalproperties,suchasthemolecularvolume.Likewise,theidentificationofa
particularodorantisnotduetotheactivationofonesinglereceptorbuttotheactivationofagroup
ofreceptorswithaspecialpatternwhichistypicalforcombinatorialcoding.Ithasbeenshownthat
receptorscanbebroadlytunedandrespondtomanydifferentodorants,beingmostresponsiveto
structurallysimilarodorants,ornarrowlytuned,respondingtoasmallgroupofodorants,as
depictedinFigure8[23].

Figure8.Olfactoryreceptor(OR)responsestoodorants.Humanshaveapproximatively396
olfactoryreceptors(OR1,OR2,OR3,etc.).Asingleolfactoryreceptorisabletorecognizedifferent
odorantmolecules.Asanexample,OR1isabletorecognizemoleculesA,B,andC(d).The
identificationofaparticularodorantiscausedbytheactivationofagroupofreceptorswithaspecial
pattern(a,b,andc).Forexample,odorantAisrecognizedbyOR1andOR2asabananaflavor(a).
MoleculeBis,inturn,recognizedbyOR1andOR3andidentifiedasanorangeflavor(b).Also,two
distinctmoleculescanberecognizedbythesamereceptorsandidentifiedashavingthesameodor(a
andc).Indeed,odorantCisalsorecognizedbyOR1andOR2asabananaflavor(c).
Asacomplement,higherconcentrationsofodorantsactivateagreaternumberofreceptors.
Hence,thenumberofactivatedreceptorsislinkedbothtotheidentityofanodorantaswellasits
concentration[39].Anotherconcernistemporality.Anodorantcancausealong‐lastingresponsein
someolfactoryreceptorsandashorterresponseinothers.Asacorollary,anindividualolfactory
receptorcangiveashortanswertosomemoleculesandalongonetoothercompounds[39].
Furthermore,theactivationofolfactoryreceptorsisnottheonlyregulationpathwayinolfaction,as
someolfactoryreceptorscanalsobeinhibitedbysomeodorants[39].SinceagivenORNexpresses
onlyoneclassofodorantreceptorinhumans,andasneuronscontainingthesametypeofreceptor
sendtheiraxonstooneorafewglomerulioftheolfactorybulb,thestudyoftheinteractionbetween
areceptorandanodorantmoleculeisakeypointinolfactionresearch[73].
Figure 8.
Olfactory receptor (OR) responses to odorants. Humans have approximatively 396 olfactory
receptors (OR 1, OR 2, OR 3, etc.). A single olfactory receptor is able to recognize different odorant
molecules. As an example, OR 1 is able to recognize molecules A, B, and C (d). The identification of a
particular odorant is caused by the activation of a group of receptors with a special pattern (a, b, and c).
For example, odorant A is recognized by OR 1 and OR 2 as a banana flavor (
a
). Molecule B is, in turn,
recognized by OR 1 and OR 3 and identified as an orange flavor (
b
). Also, two distinct molecules can
be recognized by the same receptors and identified as having the same odor (a and c). Indeed, odorant
C is also recognized by OR 1 and OR 2 as a banana flavor (c).
As a complement, higher concentrations of odorants activate a greater number of receptors.
Hence, the number of activated receptors is linked both to the identity of an odorant as well as its
concentration [
39
]. Another concern is temporality. An odorant can cause a long-lasting response in
some olfactory receptors and a shorter response in others. As a corollary, an individual olfactory receptor
can give a short answer to some molecules and a long one to other compounds [
39
]. Furthermore,
the activation of olfactory receptors is not the only regulation pathway in olfaction, as some olfactory
receptors can also be inhibited by some odorants [
39
]. Since a given ORN expresses only one class of
odorant receptor in humans, and as neurons containing the same type of receptor send their axons to
one or a few glomeruli of the olfactory bulb, the study of the interaction between a receptor and an
odorant molecule is a key point in olfaction research [73].
Sequences of thousands of receptors have been discovered from genomes. ORs, representing
more than 2% of the human genome and 4% of our proteome, belong to the typical class A GPCR
motifs. Nevertheless, ORs show a low sequence identity with nonolfactory proteins of this class, as OR
conserved motifs are either dissimilar to or more complex than those present in sequences of other
class A GPCR. Since no 3D structure of ORs either in the active or inactive state has been resolved
to date, molecular modeling approaches (homology modeling, docking, and molecular dynamics,
notably) coupled to sequence conservation and site-directed mutagenesis have helped to predict 3D
models of ORs and their binding to diverse ligands. The binding pocket is notably composed of
hydrophobic transmembrane (TM) residues from TM3, 5, 6, and 7 [
74
–
78
], such as Phe 104, that could
contribute to stabilizing the ligand through an interaction between the aromatic cycle and double
bonds from odorant molecules [
74
]. The activation mechanisms are still poorly understood since the
highly conserved motifs, critical for the activation of nonolfactory GPCRs, are not present. Some of
these mechanisms’ aspects have been recently elucidated by using molecular dynamics and point
mutations, involving the highly conserved Y252 in helix 6 and identifying the ionic lock residues
between TM3 and TM6, which usually induce conformational change and subsequent activation [
74
].
Even though these models have substantially advanced our understanding of the recognition and
activation of ORs, the structural elucidation of olfactory GPCRs would provide a great advance in the
field of olfaction, as was in the case of rhodopsin, the first G-coupled receptor to be crystallized, which
helped us to better understand light perception [39].

Int. J. Mol. Sci. 2019,20, 3018 10 of 16
In mammals, olfactory receptors are not restricted to odorant receptors of the main olfactory
epithelium. In mice, a commonly used model, the olfactory system contains different chemosensory
subsystems counting the following olfactory receptors: odorant receptors, vomeronasal receptors, trace
amine-associated receptors, formyl peptide receptors, and guanylyl cyclase receptors, most of which
belong to the GPCR class. The relationship between the main olfactory system and accessory olfactory
systems is complex and needs further investigation, especially for species other than mouse [2,23,79].
The vomeronasal organ is also present in insects, where it allows, along with antennas, the perception
of pheromones, for example. Those molecules are semiochemicals known to influence communication
in all living organisms [
80
]. According to their function, pheromones play key roles in animal alarm,
releaser, territorial information, and sex [
3
,
81
]. In humans, the vomeronasal organ is vestigial and
nonfunctional [
81
], and the existence of functional accessory olfactory systems is still unclear [
41
].
Furthermore, olfactory receptors have been described in nontypical olfactory tissues such as myocardial
and erythroid cells, ganglia, spleen, colon, prostate, and testis [
82
]. Olfactory receptors were found in
mature sperm, revealing a potential role in sperm–oocyte chemiotaxis [82].
3. New Trends in Olfaction Studies
3.1. Importance and Diversity of Human Olfaction
Since the 19th century and Broca’s claim that humans are “nonsmellers”, the belief that humans
have an underdeveloped sense of smell compared with other mammals is deeply rooted in the scientific
community [
72
]. As a matter of fact, the American Medical Association has stated that loss of vision
and audition resulted, respectively, in a disability rate of 85% and 35%, whereas it was only 3% for
olfaction loss [
41
]. Recent studies contradict this 19th century myth, revealing that humans have both
very good capacities for detecting and discriminating different odors [
72
,
83
]. The importance of odors
and their influence on mood, cognition, and behavior is now recognized [84].
Differences between individuals in perceiving odorants have received more attention recently.
Human olfaction perception is extremely variable for both specific sensitivity and general olfactory
acuity, with sensitivity varying several orders of magnitude between individuals [
85
]. The
polymorphism of olfactory receptors constitutes the molecular basis of such individual differences [
85
].
Specific anosmia, specific hyposmia (reduced ability to detect an odor), and the less-studied specific
hyperosmia (increased olfactory acuity) occurrences are extremely numerous, which are, from a
statistical point of view, more the rule than the exception [
85
,
86
]. The first logical explanation of such
large olfactory perception diversity is genetic. Among the olfactory receptor genes and pseudogenes
of the human genome, more than 60% present damaging single-nucleotide polymorphisms (SNPs) in
their coding sequence. SNPs are a major source of human genetic variability and, when they are present
in a coding sequence of the genome, they cause segregation between intact and disrupted alleles in the
population. Even though a significant concordance can be observed for several odorant thresholds,
it is suggested that olfactory acuity is a complex trait due to olfactory receptor variations as well as
differences in the downstream treatment of the information in the signaling pathway [
85
,
87
]. The
complex origin of anosmia, which is not only due to olfactory receptor genetic variations, is supported
by the observation that training can be used to improve sensitivity towards specific odorants which
were initially undetectable [
86
]. The plasticity of the olfactory system in adults has been shown by
successfully training volunteers to discriminate between odorant enantiomers [88].
Olfaction studies are no longer restricted to perfumers and flavorists, as illustrated by military
interest in the field [
89
]. The importance of the olfactory sense is now gaining the attention of the
medical domain because an impaired olfactory system impacts physical health, nutrition and eating
pleasure, and more generally, the quality of life [
84
,
90
]. Each of the four steps of the olfaction process
can be affected: nasal airflow can be impacted by nasal polyposis, chronic rhinitis can result in
peri-receptor event alteration, transduction via the primary olfactory neurons can be impaired by viral
rhinitis, and neurodegenerative diseases are mainly responsible for brain processing problems [
40
,
90
].

Int. J. Mol. Sci. 2019,20, 3018 11 of 16
Raising awareness of olfaction’s importance in human health is crucial. Smell loss can be an early sign
of neurodegenerative diseases such as Alzheimer’s disease or Parkinson’s disease and can assist in
their diagnosis [90].
3.2. Electronic Nose and Machine Learning
High interindividual variations in olfaction capacities and increasing sets of data pave the way for
new fields in olfaction studies, such as the electronic nose and machine learning.
Electronic noses are not new; the first one was designed four decades ago and consisted of
simple gas detectors, the ratio of the signals from the different transducers being processed to identify
an odor [
91
]. Electronic noses are made up of an array of different gas sensors that can interact
with volatile molecules and produce a signal, such as an electrical or optical signal, that can be
amplified and finally processed [
92
]. Progress in electronic nose technology is due to concomitant
improvements of sensor properties, signal preparation and pattern recognition, as reviewed in [
93
].
To cope with the high sensitivity, high selectivity, and the number of odorants potentially detected
by the human nose, numerous studies have already been undertaken, but many more must still be
performed before designing an instrument similar to the human olfactory system that can work in
various environmental conditions [
94
]. It remains true that some volatile molecules can be detected
by artificial olfaction machines but not by the human nose. The diversity of detected odors can be
improved by increasing the number of specific sensing elements, even though appropriate sensitivity
is difficult to reach, as human nose detection limits are at least 3–5 orders of magnitude lower than
those of actual gas detectors [
92
]. The selectivity is linked to the type of sensing element used. Metal
oxides and conducting polymers have poor selectivity [
92
], while semiconducting oxide nanobelts are
more stable and highly sensitive [
95
]. Chiral polymer composite allows for the differential detection
of enantiomeric volatile compounds [
96
]. More recently, it was proposed to use olfactory receptor
proteins as sensing elements, but as they are transmembrane proteins, it was impossible to obtain the
required stability. Odorant-binding proteins are more appropriate candidates as they are soluble and
extremely stable with temperature and in organic solvents [
94
]. The main problems that remain to be
solved are the need for an aqueous interface and the low dissociation kinetics which are incompatible
with continuous monitoring of odorants. Design of site-specific mutated proteins with modified
ligand-binding specificity could be a solution [
94
]. Although an instrument capable of discriminating
odors with human-like performance still needs to be designed, the application domains are increasing
with technological improvements in the military field [
89
], environmental monitoring [
97
], medical
diagnostics [98], and food and beverage quality control [99–101].
Questions still remain regarding the multifaceted olfactory code and the complex and
high-dimensional aspects of data from olfactory research. During the last two decades, artificial neural
networks were used to help elucidate structure–odor relationships: prediction of the camphoraceous
or fruity odor of aliphatic alcohols [
102
], the musk odor of tetralin and indan compounds [
103
], the
aroma quality and threshold values of pyrazines [
104
], and so forth. More recently, machine learning
was used to facilitate complex data-driven research in the olfactory field; for example, to predict the
response of olfactory receptors [
105
]. Machine learning methods are increasingly widely used in
human olfactory research at different levels: physiology of odor detection and recognition, olfactory
acuity as diagnostic biomarkers, and odor recognition from physicochemical properties of volatiles
(for review, see [106]).
4. Conclusions
Determining whether it is possible to predict the odor of a molecule based on its structure is
certainly not trivial. The literature is replete with descriptions of odor and threshold values for
thousands of natural and synthetic odorants, which can be useful for anyone wishing to know what
smell is associated with an already-known molecule. Structural similarity can be used in some cases to
predict the odor of a particular compound but, as mentioned before, it is far from being 100% reliable.

Int. J. Mol. Sci. 2019,20, 3018 12 of 16
Recent progress in chemistry, physiology, neurobiology, and genetics with the help of
bioinformatics and machine learning should allow the better understanding of the human olfaction
system. Studying olfactory receptor–odorant interactions and the activation mechanisms is certainly
the most attractive approach in order to predict an odor based on molecular structure. Nowadays,
it seems impossible to test every single olfactory receptor with every possible odorant, but with the
help of bioinformatics approaches, this goal is perhaps reachable, especially if the 3D structure of
ORs were found. The combinatorial coding scheme of odorants requires a better understanding of
how the combination of OR activation is processed in the brain to link odor to olfactory perception.
Taking into account the enormous datasets generated by combinatorial coding, machine learning can
be helpful. Intrinsic signal imaging of the olfactory bulb has been suggested by some authors, but
technical improvements are still required. Depending on the goal of each study, a global approach may
be appropriate. For example, when the main question is not knowing how the olfactory system works
but rather whether the consumer will enjoy the food/perfume for commercial applications, trained
panelists would be the best solution to provide sensory evaluations of the products and to reflect
individual variability [107,108].
In many traditions, sniffing food is considered impolite. This cultural issue has gradually
contributed to the decline of olfaction in everyday life. From a public health point of view, it is crucial
to restore the pivotal role of olfaction among our five senses by changing mentalities and encouraging
children to smell and enjoy what they eat. Elders should also stimulate their olfaction to counteract their
loss in olfactory ability, since this is linked to a loss of appetite, which can have dramatic consequences
on health and even be a marker of neurodegenerative disorders.
Funding:
This work has been supported by the Education, Audiovisual and Culture Executive Agency (EACEA),
through the EOHUB project (600873-EPP-1-2018-1ES-EPPKA2-KA).
Acknowledgments:
M.D. and L.L. thank the Fonds National de la Recherche Scientifique from Belgium
(FRS-FNRS).
Conflicts of Interest: The authors declare no conflict of interests.
References
1.
Depelteau, J.S.; Brenzinger, S.; Briegel, A. Bacterial and Archaeal Cell Structure. In Reference Module in Life
Sciences; Elsevier: Amsterdam, The Netherlands, 2019.
2. Grabe, V.; Sachse, S. Fundamental principles of the olfactory code. Biosystems 2018,164, 94–101. [CrossRef]
[PubMed]
3.
Choi, N.-E.; Han, J.H. How Flavor Works—The Science of Taste and Aroma; John Wiley & Sons, Ltd.: Chichester,
UK, 2015.
4. Padodara, R.J.; Jacob, N. Olfactory Sense in Different Animals. Indian J. Vet. Sci. 2014,2, 1–14.
5.
Calvo-Ochoa, E.; Byrd-Jacobs, C. The Olfactory System of Zebrafish as a Model for the Study of Neurotoxicity
and Injury: Implications for Neuroplasticity and Disease. Int. J. Mol. Sci.
2019
,20, 1639. [CrossRef] [PubMed]
6.
Wang, Z.; Zhou, Y.; Luo, Y.; Zhang, J.; Zhai, Y.; Yang, D.; Zhang, Z.; Li, Y.; Storm, D.; Ma, R. Gene Expression
Profiles of Main Olfactory Epithelium in Adenylyl Cyclase 3 Knockout Mice. Int. J. Mol. Sci.
2015
,16,
28320–28333. [CrossRef] [PubMed]
7.
Raji, J.I.; DeGennaro, M. Genetic analysis of mosquito detection of humans. Curr. Opin. Insect Sci.
2017
,20,
34–38. [CrossRef] [PubMed]
8.
Reed, D.R.; Knaapila, A. Genetics of Taste and Smell: Poisons and Pleasures. NIH Public Access
2010
,94,
213–240.
9.
Vaglio, S. Chemical communication and mother-infant recognition. Commun. Integr. Biol.
2009
,2, 279–281.
[CrossRef]
10.
Sparkes, A.C. Ethnography and the senses: Challenges and possibilities. Qual. Res. Sport Exerc.
2009
,1,
21–35. [CrossRef]
11.
Sarafoleanu, C.; Mella, C.; Georgescu, M.; Perederco, C. The importance of the olfactory sense in the human
behavior and evolution. J. Med. Life 2009,2, 196–198.

Int. J. Mol. Sci. 2019,20, 3018 13 of 16
12.
Ferdenzi, C.; Joussain, P.; Digard, B.; Luneau, L.; Djordjevic, J.; Bensafi, M. Individual Differences in Verbal
and Non-Verbal Affective Responses to Smells: Influence of Odor Label Across Cultures. Chem. Senses
2016
,
42, 37–46. [CrossRef]
13.
Hummel, T.; Kobal, G.; Gudziol, H.; Mackay-Sim, A. Normative data for the “Sniffin’ Sticks” including tests
of odor identification, odor discrimination, and olfactory thresholds: An upgrade based on a group of more
than 3,000 subjects. Eur. Arch. Oto-Rhino-Laryngology 2007,264, 237–243. [CrossRef] [PubMed]
14.
Sorokowska, A.; Hummel, T. Polska wersja testu Sniffin’ Sticks—Adaptacja i normalizacja. Otolaryngol. Pol.
2014,68, 308–314. [CrossRef] [PubMed]
15.
Sorokowska, A.; Schriever, V.A.; Gudziol, V.; Hummel, C.; Hähner, A.; Iannilli, E.; Sinding, C.; Aziz, M.;
Seo, H.S.; Negoias, S.; et al. Changes of olfactory abilities in relation to age: Odor identification in more than
1400 people aged 4 to 80 years. Eur. Arch. Oto-Rhino-Laryngology
2015
,272, 1937–1944. [CrossRef] [PubMed]
16.
Doty, R.; Shaman, P.; Applebaum, S.; Giberson, R.; Siksorski, L.; Rosenberg, L. Smell identification ability:
Changes with age. Science 1984,226, 1441–1443. [CrossRef] [PubMed]
17.
Sorokowski, P.; Karwowski, M.; Misiak, M.; Marczak, M.K.; Dziekan, M.; Hummel, T.; Sorokowska, A. Sex
differences in human olfaction: A meta-analysis. Front. Psychol. 2019,10, 242. [CrossRef] [PubMed]
18.
Buck, L.; Axel, R. A novel multigene family may encode odorant receptors: A molecular basis for odor
recognition. Cell 1991,65, 175–187. [CrossRef]
19. John, P.J. Basics of Biopsychology, 7th ed.; Allyn & Bacon, Inc.: London, UK, 2007.
20.
Herrmann, A. The Chemistry and Biology of Volatiles; Herrmann, A., Ed.; John Wiley & Sons, Ltd.: Chichester,
UK, 2010; p. 428.
21.
De Gennaro, G.; Farella, G.; Marzocca, A.; Mazzone, A.; Tutino, M. Indoor and outdoor monitoring of volatile
organic compounds in school buildings: Indicators based on health risk assessment to single out critical
issues. Int. J. Environ. Res. Public Health 2013,10, 6273–6291. [CrossRef] [PubMed]
22.
Yoshii, F. Structure-odor relations: A modern perspective. In Handbook of Olfaction and Gustation; Doty, R.L.,
Ed.; Taylor & Francis Group: Abingdon, UK, 2003; pp. 457–492.
23.
Nara, K.; Saraiva, L.R.; Ye, X.; Buck, L.B. A Large-Scale Analysis of Odor Coding in the Olfactory Epithelium.
J. Neurosci. 2011,31, 9179–9191. [CrossRef]
24.
Dufoss
é
, L.; Latrasse, A.; Spinnler, H.-E. Importance des lactones dans les ar
ô
mes alimentaires. Sci des
Aliment. 1994,14, 17–50.
25. Goldstein, N. Getting to know the odor compounds. Biocycle 2002,43, 42–44.
26.
Buettner, A. Springer Handbook of Odor; Buettner, A., Ed.; Springer International Publishing: Cham,
Switzerland, 2017.
27.
Laing, D.G. Relationship between Molecular Structure, Concentration and Odor Qualities of Oxygenated
Aliphatic Molecules. Chem. Senses 2003,28, 57–69. [CrossRef] [PubMed]
28.
Poivet, E.; Tahirova, N.; Peterlin, Z.; Xu, L.; Zou, D.-J.; Acree, T.; Firestein, S. Functional odor classification
through a medicinal chemistry approach. Sci. Adv. 2018,4, eaao6086. [CrossRef] [PubMed]
29.
Zarzo, M. Effect of Functional Group and Carbon Chain Length on the Odor Detection Threshold of Aliphatic
Compounds. Sensors 2012,12, 4105–4112. [CrossRef] [PubMed]
30. Nobuhara, A. Syntheses of Unsaturated Lactones. Agric. Biol. Chem. 1968,32, 1016–1020.
31.
Brookes, J.C.; Horsfield, A.P.; Stoneham, A.M. Odour character differences for enantiomers correlate with
molecular flexibility. J. R. Soc. Interface 2009,6, 75–86. [CrossRef] [PubMed]
32.
Boelens, M.H.; van Gemert, L.J. Volatile character-impact sulfur compounds and their sensory properties.
Perfum. Flavorist 1993,18, 29–39.
33.
Meierhenrich, U.J.; Golebiowski, J.; Fernandez, X.; Cabrol-Bass, D. De la mol
é
cule
à
l’odeur: Les bases
moléculaires des premières étapes de l’olfaction. L. Actual. Chim. 2005, 29–40.
34. Rossiter, K.J. Structure-odor relationships. Chem. Rev. 1996,96, 3201–3240. [CrossRef]
35.
Kraft, P.; Bajgrowicz, J.A.; Denis, C.; Frater, G. Odds and Trends: Recent Developments in the Chemistry of
Odorants. Angew. Chemie Int. Ed. 2000,39, 2980–3010. [CrossRef]
36.
Brenna, E.; Fuganti, C.; Serra, S. Enantioselective perception of chiral odorants. Tetrahedron: Asymmetry
2003
,
14, 1–42. [CrossRef]
37.
Sell, C.S. On the unpredictability of odor. Angew. Chemie Int. Ed.
2006
,45, 6254–6261. [CrossRef] [PubMed]
38.
Ohloff, G.; Maurer, B.; Winter, B.; Giersch, W. Structural and Configurational Dependence of the Sensory
Process in Steroids. Org. Biol. Chemie Chim. Org. Chim. Biol. 1983,66, 192–217. [CrossRef]

Int. J. Mol. Sci. 2019,20, 3018 14 of 16
39.
Su, C.-Y.; Menuz, K.; Carlson, J.R. Olfactory Perception: Receptors, Cells, and Circuits. Cell
2009
,139, 45–59.
[CrossRef] [PubMed]
40. Bonfils, P. Odorat: De l’aéroportage au cortex. Bull. Acad. Natl. Med. 2014,198, 1109–1120. [PubMed]
41.
Sela, L.; Sobel, N. Human olfaction: A constant state of change-blindness. Exp. Brain Res.
2010
,205, 13–29.
[CrossRef] [PubMed]
42.
Maresh, A.; Rodriguez Gil, D.; Whitman, M.C.; Greer, C.A. Principles of glomerular organization in the
human olfactory bulb—Implications for odor processing. PLoS ONE 2008,3. [CrossRef] [PubMed]
43. Herden, G. Some aspects of qualitative data analysis. Math. Soc. Sci. 1993,26, 105–138. [CrossRef]
44.
Boelens, M.H.; Boelens, H.; Van Gemert, L.J. Sensory Properties of Optical Isomers. Perfum. Flavonist
1993
,
18, 1.
45.
Chastrette, M. Classification of Odors and Structure–Odor Relationships. In Olfaction, Taste, and Cognition;
Rouby, C., Schaal, B., Dubois, D., Gervais, R., Holley, A., Eds.; Cambridge University Press: Cambridge, UK,
2009; pp. 100–116.
46.
Ohloff, G.; Winter, B.; Fehr, C. Chemical Classification and Structure—Odour Relationships. In Perfumes;
Springer Netherlands: Dordrecht, Netherlands, 1994; pp. 287–330.
47.
Jordan, R.; Kollo, M.; Schaefer, A.T. Sniffing Fast: Paradoxical Effects on Odor Concentration Discrimination
at the Levels of Olfactory Bulb Output and Behavior. Eneuro 2018,5. [CrossRef]
48.
Wilkes, F.J.; Laing, D.G.; Hutchinson, I.; Jinks, A.L.; Monteleone, E. Temporal processing of olfactory stimuli
during retronasal perception. Behav. Brain Res. 2009,200, 68–75. [CrossRef]
49.
Nagashima, A.; Touhara, K. Enzymatic Conversion of Odorants in Nasal Mucus Affects Olfactory Glomerular
Activation Patterns and Odor Perception. J. Neurosci. 2010,30, 16391–16398. [CrossRef] [PubMed]
50.
Thiebaud, N.; da Silva, S.V.; Jakob, I.; Sicard, G.; Chevalier, J.; M
é
n
é
trier, F.; Berdeaux, O.; Artur, Y.;
Heydel, J.M.; Le Bon, A.M. Odorant Metabolism Catalyzed by Olfactory Mucosal Enzymes Influences
Peripheral Olfactory Responses in Rats. PLoS ONE 2013,8, 34–56. [CrossRef] [PubMed]
51. Pelosi, P. Odorant-binding proteins. Crit. Rev. Biochem. Mol. Biol. 1994,29, 199–228. [CrossRef] [PubMed]
52.
Tegoni, M.; Pelosi, P.; Vincent, F.; Spinelli, S.; Campanacci, V.; Grolli, S.; Ramoni, R.; Cambillau, C. Mammalian
odorant binding proteins. Biochim. Biophys. Acta Protein Struct. Mol. Enzymol.
2000
,1482, 229–240. [CrossRef]
53.
Homma, R.; Lv, X.; Sato, T.; Imamura, F.; Zeng, S.; Nagayama, S. Narrowly Confined and Glomerulus-Specific
Onset Latencies of Odor-Evoked Calcium Transients in the Juxtaglomerular Cells of the Mouse Main
Olfactory Bulb. Eneuro 2019,6. [CrossRef] [PubMed]
54.
Kay, L.M.; Sherman, S.M. An argument for an olfactory thalamus. Trends Neurosci.
2007
,30, 47–53. [CrossRef]
55.
Geramita, M.; Urban, N.N. Differences in Glomerular-Layer-Mediated Feedforward Inhibition onto Mitral
and Tufted Cells Lead to Distinct Modes of Intensity Coding. J. Neurosci. 2017,37, 1428–1438. [CrossRef]
56.
Bizer, A.; Parabucki, A.; Smear, M.; Munoz, A.E.; Shusterman, R.; Bala, A.D.S.; Morris, G. Odor concentration
change coding in the olfactory bulb. Eneuro 2019,6. [CrossRef]
57.
Jaubert, J.N.; Tapiero, C.; Dore, J.C. The field of odours: Toward a universal language for odour relationships.
Perfum. Flavorist 1995,20, 1–16.
58.
David, S. Linguistic Expressions for Odors in French. In Olfaction, Taste, and Cognition; Rouby, C., Schaal, B.,
Dubois, D., Gervais, R., Holley, A., Eds.; Cambridge University Press: Cambridge, UK, 2009; pp. 82–99.
59.
Block, E.; Jang, S.; Matsunami, H.; Sekharan, S.; Dethier, B.; Ertem, M.Z.; Gundala, S.; Pan, Y.; Li, S.; Li, Z.; et al.
Implausibility of the vibrational theory of olfaction. Proc. Natl. Acad. Sci. 2015,112, 2766–2774. [CrossRef]
60. Malcolm Dyson, G. The scientific basis of odour. J. Soc. Chem. Ind. 1938,57, 647–651. [CrossRef]
61. De Jong, H.G.B.S. G.G.. model for the stimulation of organ of smell. Proc. Acad. Sci. 1937,40, 302–306.
62.
Davies, J.T.; Taylor, F.H. A Model System for the Olfactory Membrane. Nature
1954
,174, 693–694. [CrossRef]
[PubMed]
63. Briggs, M.H.; Duncan, R.B. Odor receptors. Nature 1961,91, 1310–1311. [CrossRef] [PubMed]
64.
Rosenberg, B.; Misra, T.N.; Switzer, R. Mechanism of Olfactory Transduction. Nature
1968
,217, 423–427.
[CrossRef]
65.
Mozell, M.M. Evidence for a Chromatographic Model of Olfaction. J. Gen. Physiol.
2004
,56, 46–63. [CrossRef]
66. Amoore, J.E. Specific anosmia and the concept of primary odors. Chem. Senses 1977,2, 267–281. [CrossRef]
67.
Triller, A.; Boulden, E.A.; Churchill, A.; Hatt, H.; Englund, J.; Spehr, M.; Sell, C.S. Odorant–Receptor
Interactions and Odor Percept: A Chemical Perspective. Chem. Biodivers. 2008,5, 862–886. [CrossRef]
68. Amoore, J.E. Stereochemical theory of olfaction. Nature 1963,198, 271–272. [CrossRef]

Int. J. Mol. Sci. 2019,20, 3018 15 of 16
69.
Silva Teixeira, C.S.; Cerqueira, N.M.F.S.A.; Silva Ferreira, A.C. Unravelling the Olfactory Sense: From the
Gene to Odor Perception. Chem. Senses 2015,41, 105–121. [CrossRef]
70. DeMaria, S.; Ngai, J. The cell biology of smell. J. Cell Biol. 2010,191, 443–452. [CrossRef] [PubMed]
71.
Block, E. Molecular Basis of Mammalian Odor Discrimination: A Status Report. J. Agric. Food Chem.
2018
,66,
13346–13366. [CrossRef] [PubMed]
72.
McGann, J.P. Poor human olfaction is a 19th-century myth. Science
2017
,356, eaam7263. [CrossRef] [PubMed]
73.
Malnic, B. Searching for the ligands of odorant receptors. Mol. Neurobiol.
2007
,35, 175–181. [CrossRef]
[PubMed]
74.
De March, C.A.; Yu, Y.; Ni, M.J.; Adipietro, K.A.; Matsunami, H.; Ma, M.; Golebiowski, J. Conserved Residues
Control Activation of Mammalian G Protein- Coupled Odorant Receptors. J. Am. Chem. Soc.
2015
,137,
8611–8616. [CrossRef] [PubMed]
75.
De March, C.A.; Kim, S.; Antonczak, S.; Goddard, W.A., III; Golebiowski, J. G protein-coupled odorant
receptors: From sequence to structure. Protein Sci. 2015,24, 1543–1548. [CrossRef] [PubMed]
76.
Baud, O.; Etter, S.; Spreafico, M.; Bordoli, L.; Schwede, T.; Vogel, H.; Pick, H. The Mouse Eugenol Odorant
Receptor: Structural and Functional Plasticity of a Broadly Tuned Odorant Binding Pocket. Biochemistry
2011,50, 843–853. [CrossRef] [PubMed]
77.
Man, O.; Gilad, Y.; Lancet, D. Prediction of the odorant binding site of olfactory receptor proteins by human –
mouse comparisons. Protein Sci. 2004,13, 240–254. [CrossRef] [PubMed]
78.
Floriano, W.B.; Vaidehi, N.; Goddard, W.A. Making sense of olfaction through predictions of the 3-D structure
and function of olfactory receptors. Chem. Senses 2004,29, 269–290. [CrossRef] [PubMed]
79.
Mohrhardt, J.; Nagel, M.; Fleck, D.; Ben-Shaul, Y.; Spehr, M. Signal detection and coding in the accessory
olfactory system. Chem. Senses 2018,43, 667–695. [CrossRef] [PubMed]
80.
Brennan, P.A. Pheromones and Mammalian Behavior. In The Neurobiology of Olfaction; Menini, A., Ed.; CRC
Press/Taylor & Francis: Abingdon, UK, 2010.
81.
Gomez-Diaz, C.; Benton, R. The joy of sex pheromones. EMBO Rep.
2013
,14, 874–883. [CrossRef] [PubMed]
82.
Milardi, D.; Colussi, C.; Grande, G.; Vincenzoni, F.; Pierconti, F.; Mancini, F.; Baroni, S.; Castagnola, M.;
Marana, R.; Pontecorvi, A. Olfactory receptors in semen and in the male tract: From proteome to proteins.
Front. Endocrinol. 2018,8, 379. [CrossRef] [PubMed]
83.
Olofsson, J.K.; Wilson, D.A. Human Olfaction: It Takes Two Villages. Curr. Biol.
2018
,28, 108–110. [CrossRef]
[PubMed]
84.
Soudry, Y.; Lemogne, C.; Malinvaud, D.; Consoli, S.M.; Bonfils, P. Olfactory system and emotion: Common
substrates. Eur. Ann. Otorhinolaryngol. Head Neck Dis. 2011,128, 18–23. [CrossRef] [PubMed]
85.
Menashe, I.; Abaffy, T.; Hasin, Y.; Goshen, S.; Yahalom, V.; Luetje, C.W.; Lancet, D. Genetic Elucidation of
Human Hyperosmia to Isovaleric Acid. PLoS Biol. 2007,5. [CrossRef] [PubMed]
86.
Croy, I.; Olgun, S.; Mueller, L.; Schmidt, A.; Muench, M.; Hummel, C.; Gisselmann, G.; Hatt, H.; Hummel, T.
Peripheral adaptive filtering in human olfaction? Three studies on prevalence and effects of olfactory training
in specific anosmia in more than 1600 participants. Cortex 2015,73, 180–187. [CrossRef]
87.
Trimmer, C.; Keller, A.; Murphy, N.R.; Snyder, L.L.; Willer, J.R.; Nagai, M.H.; Katsanis, N.; Vosshall, L.B.;
Matsunami, H.; Mainland, J.D. Genetic variation across the human olfactory receptor repertoire alters odor
perception. Proc. Natl. Acad. Sci. 2019,16, 9475–9480. [CrossRef]
88.
Feng, G.; Zhou, W. Nostril-specific and structure-based olfactory learning of chiral discrimination in human
adults. Elife 2019,8, e41296. [CrossRef]
89.
Nagappan, P.G.; Subramaniam, S.; Wang, D.Y. Olfaction as a soldier– a review of the physiology and its
present and future use in the military. Mil. Med. Res. 2017,4, 9. [CrossRef]
90.
Doty, R.L.; Kamath, V. The influences of age on olfaction: A review. Front. Psychol.
2014
,5, 1–20. [CrossRef]
91.
Persaud, K.; Dodd, G. Analysis of discrimination mechanisms in the mammalian olfactory system using a
model nose. Nature 1982,299, 352–355. [CrossRef] [PubMed]
92.
Pelosi, P.; Zhu, J.; Knoll, W. From Gas Sensors to Biomimetic Artificial Noses. Chemosensors
2018
,6, 32.
[CrossRef]
93.
Gardner, J.W.; Bartlett, P.N. A brief history of electronic noses. Sens. Actuators B Chem.
1994
,18, 210–211.
[CrossRef]
94.
Pelosi, P.; Zhu, J.; Knoll, W. Odorant-binding proteins as sensing elements for odour monitoring. Sensors
2018,18, 3248. [CrossRef] [PubMed]

Int. J. Mol. Sci. 2019,20, 3018 16 of 16
95.
Comini, E.; Faglia, G.; Sberveglieri, G.; Pan, Z.; Wang, Z.L. Stable and highly sensitive gas sensors based on
semiconducting oxide nanobelts. Appl. Phys. Lett. 2002,81, 1869–1871. [CrossRef]
96.
Severin, E.J.; Sanner, R.D.; Doleman, B.J.; Lewis, N.S. Differential Detection of Enantiomeric Gaseous Analytes
Using Carbon Black
−
Chiral Polymer Composite, Chemically Sensitive Resistors. Anal. Chem.
1998
,70,
1440–1443. [CrossRef] [PubMed]
97.
Wilson, A.D. Review of Electronic-nose Technologies and Algorithms to Detect Hazardous Chemicals in the
Environment. Procedia Technol. 2012,1, 453–463. [CrossRef]
98.
Wojnowski, W.; Dymerski, T.; G ˛ebicki, J.; Namie´snik, J. Electronic Noses in Medical Diagnostics. Curr. Med.
Chem. 2019,26, 197–215. [CrossRef] [PubMed]
99.
Loutfi, A.; Coradeschi, S.; Mani, G.K.; Shankar, P.; Rayappan, J.B.B. Electronic noses for food quality: A review.
J. Food Eng. 2015,144, 103–111. [CrossRef]
100.
Zou, G.; Xiao, Y.; Wang, M.; Zhang, H. Detection of bitterness and astringency of green tea with different
taste by electronic nose and tongue. PLoS ONE 2018,13, e0206517. [CrossRef] [PubMed]
101.
Long, Q.; Li, Z.; Han, B.; Gholam Hosseini, H.; Zhou, H.; Wang, S.; Luo, D. Discrimination of Two Cultivars of
Alpinia Officinarum Hance Using an Electronic Nose and Gas Chromatography-Mass Spectrometry Coupled
with Chemometrics. Sensors 2019,19, 572. [CrossRef] [PubMed]
102.
Chastrette, M.; Cretin, D.; El Aïdi, C. Structure
−
Odor Relationships: Using Neural Networks in the Estimation
of Camphoraceous or Fruity Odors and Olfactory Thresholds of Aliphatic Alcohols. J. Chem. Inf. Comput.
Sci. 1996,36, 108–113. [CrossRef] [PubMed]
103.
Cherqaoui, D.; Esseffar, M.; Villemin, D.; Cense, J.-M.; Chastrette, M.; Zakarya, D. Structure-musk odour
relationship studies of tetralin and indan compounds using neural networks. New J. Chem.
1998
,22, 839–843.
[CrossRef]
104.
Wailzer, B.; Klocker, J.; Buchbauer, G.; Ecker, G.; Wolschann, P. Prediction of the Aroma Quality and the
Threshold Values of Some Pyrazines Using Artificial Neural Networks. J. Med. Chem.
2001
,44, 2805–2813.
[CrossRef] [PubMed]
105.
Gabler, S.; Soelter, J.; Hussain, T.; Sachse, S.; Schmuker, M. Physicochemical vs. Vibrational Descriptors for
Prediction of Odor Receptor Responses. Mol. Inform. 2013,32, 855–865. [CrossRef] [PubMed]
106. Lötsch, J.; Kringel, D.; Hummel, T. Machine Learning in Human Olfactory Research. Chem. Senses 2019,44,
11–22. [CrossRef]
107.
Giacalone, D.; Degn, T.K.; Yang, N.; Liu, C.; Fisk, I.; Münchow, M. Common roasting defects in coffee:
Aroma composition, sensory characterization and consumer perception. Food Qual. Prefer.
2019
,71, 463–474.
[CrossRef]
108.
S
á
enz-Navajas, M.-P.; Arias, I.; Ferrero-del-Teso, S.; Fern
á
ndez-Zurbano, P.; Escudero, A.; Ferreira, V.
Chemo-sensory approach for the identification of chemical compounds driving green character in red wines.
Food Res. Int. 2018,109, 138–148. [CrossRef]
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).