scieee Science in your language
[en] (orig)

Melody discrimination and protein fold classification

Read accessible full text

Melody discrimination and protein fold classification

Author: Bywater, Robert,Middleton, Jonathan
Year: 2016
Source: https://trepo.tuni.fi/bitstream/10024/100796/1/melody_discrimination_and.pdf
Melody disc imina ion and
p o ein old classi ica ion
Robe P. Bywa e
a
, Jona han N. Middle on
b,c,
*
a
F ancis C ick Ins i u e, London NW1 1AT, UK
b
Depa men o Music, Eas e n Washing on Uni e si y, Cheney, WA 99004, USA
c
School o In o ma ion Sciences, Uni e si y o Tampe e, 33041, Finland
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (J.N. Middle on).
Abs ac
One o he g ea es challenges in heo e ical biophysics and bioin o ma ics is he
iden i ica ion o p o ein olds om sequence da a. This can be ega ded as a pa e n
ecogni ion p oblem. In his pape we epo he use o a melody gene a ion
so wa e whe e he inpu s a e de i ed om calcula ions o e olu iona y
in o ma ion, seconda y s uc u e, lexibili y, hyd opa hy and sol en accessibili y
om mul iple sequence alignmen da a. The melodies so gene a ed a e de i ed
om he sequence, and by in e ence, o he old, in ways ha gi e each old a
sound ep esen a ion ha may acili a e analysis, ecogni ion, o compa ison wi h
o he sequences.
Keywo d: Bioin o ma ics
1. In oduc ion
Globula p o eins a e linea copolyme s o amino acid esidues ha ha e s e ches
o mo e o less egula geome y which a e packed oge he in ways ha a i s
sigh appea any hing bu sys ema ic. The locally egula egions, called seconda y
s uc u es, a e ei he helical (α-o 3
10
helix ype) o β-s ands, which can a ange
hemsel es in ei he a mu ually pa allel o an ipa allel ashion o o m so-called
β-shee s. Each amino acid esidue ype has a di e en p opensi y o a ou one o
o he o hese s uc u al a angemen s [1,2,3,4]. The old ha ensues is likewise
dependen on ano he p ope y o he amino acids, he hyd opa hy, o deg ee o
Recei ed:
18 July 2016
Re ised:
4 Sep embe 2016
Accep ed:
30 Sep embe 2016
Heliyon 2 (2016) e00175
h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
which he side chain o he esidue seeks o a oids con ac wi h he sol en wa e .
This in u n a ec s he deg ee o which a gi en esidue will be bu ied in he in e io
o lie on he su ace. These a e h ee o he majo a ibu es o polypep ide
sequences ha con i e o de e mine he inal 3-dimensional (3D) s uc u e, and all
h ee can be de e mined by p esen -day bioin o ma ics me hodology (see
Me hods). One mo e complica ion ha needs o be deal wi h is he ac ha he
polypep ides hemsel es may old in o mo e han one “independen ly olding uni ”
along he chain. These uni s a e e e ed o as domains. Fo small p o eins he e is
ypically only a single domain, bu la ge p o eins may ha e wo o mo e. These
can be simila in cha ac e (and as such may be he esul o a gene duplica ion
e en ) o di e en , depending on he unc ional needs o he p o ein. The assembly
p ocess, which is well unde s ood a he le el o molecula biology, enlis s a
ca aly ic p ocess e e ed o as splicing. This in ol es selec ing disjoin sec ions
om he o iginal DNA sequence (genome) o ansla ion in o he “language”o
p o eins whe ea e he splicing p ocess akes place in a de ined o de . P o ein
olds, o mo e co ec ly, domain olds, can be classi ied in one o se e al ways [5,
6,7]. Wi h his as a backg ound o he pa e n ecogni ion p oblem, we now
desc ibe he app oach we a e aking owa ds i s solu ion.
He e, we explo e he use o melodic soni ica ion o disce n pa e ns o da a ha a e
based on ce ain ea u es associa ed wi h p o ein s uc u es: e olu iona y
in o ma ion, chemical p opensi ies and physical a ibu es. We an icipa e ha he
pe cep ion o p o ein da a ia sonic ep esen a ion can assis esea che s in he
p ocess o pa e n ecogni ion and s uc u al unde s anding. Melodic pa e ns o
sound can become ools o assis p o ein chemis s in assigning old ypes o a gi en
p o ein wi h a known sequence, bu unknown 3D s uc u e.
The po en ial bene i s o using audi o y display o scien i ic da a as a means o
de i e analy ical meaning ha e been o mally demons a ed since 1994 [8,9,10].
These au ho s a gued ha he human audi o y sys em has he sophis ica ion and he
abili y o in e p e sounds “using mul iple laye s o unde s anding”[10].In
e e ence o he choice o analy ical lis ening o e isual e iew o da a,
sugges ions ha e been made [11] o he e ec ha “pa e ns may eme ge which a e
o he wise unde ec able”. Al hough he analy ical and in e p e a i e p ocesses
( h ough lis ening) a e no comple ely unde s ood, he pe cep ual ap i udes ha e
ecognized me i s: da a- o-sound pa ame e mappings and applica ions now se e
many disciplines such as “chaos heo y, bio-medicine, in e aces o isually
disabled people, da a mining, and seismology”[10]. Mo e speci ically o genome
science, we an icipa e ha soni ica ions could highligh pa e ns o ind mu an s
and gene ic disease ma ke s o which a ich collec ion o sequence da a is
a ailable.
A icle No~e00175
2h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
While mos da a a e obse ed, soni ica ions ha a e designed speci ically o
cohe en da a- o-sound ep esen a ion can complemen mo e adi ional isual
displays o da a and o e a use ul seconda y pe cep ualiza ion o analyze da a se s.
In his pape we seek o c ea e soni ica ions o p o ein s uc u es ha can uniquely
enhance ou abili y o ecognize pa e ns ela ed o p o ein olds. We ha e chosen
pi ch as he mos basic and p ima y soni ica ion elemen o ep esen da a; hence
he da a se s a e con e ed in o sequences o pi ches. The con e sion esul s a e
melodic in na u e, demons a ing a simple and s aigh o wa d con e sion o
clea ly pe cei e he soni ied da a pa e ns. O he sound elemen s such as ha mony
and coun e poin wi h imb al a ie y did no yield simple lis ening en i onmen s
o sequen ial da a pa e n ecogni ion. In some cases hy hmic mapping was
applied as a seconda y soni ica ion pa ame e o enhance he melodies.
The p esen s udy con ibu es o a ela i ely small collec ion o p e iously explo ed
soni ica ions o p o ein s uc u es. In many cases, p e ious s udies sough o
educa e and inspi e eade s wi h in e disciplina y connec ions be ween music and
science, and making science mo e widely accessible. E o s, such as he Li e
Music p ojec , we e designed o dual pu poses o c ea ing new pe cep ual models
o p o ein analysis and enhancing elec o-acous ic music composi ions [12,13,
14,15]. In each soni ica ion s udy o p o eins, he p ojec s ha e u ilized hei own
se o mapping sys ems o con e da a- o-music. As an example, he Li e Music
p ojec [12] mapped p o ein esidues o pi ches ia hyd opa hy scales, a concep
ha is simila o one o he mapping sys ems explained la e in his pape (wi h
esul s ha a e qui e di e en ). In he Takahashi and Mille pilo s udy [13] amino
acids we e assigned o disc e e pi ches, and hen o cho ds and hy hm o make he
esul s mo e musical.
2. Me hods
2.1. Da a p epa a ion om p o ein sequence da abases
The soni ica ion s udy epo ed a his ime is ea u ing h ee p o eins, each
ep esen ing h ee SCOP [5] classes, α-helix ype, p o eins composed o β-s ands,
and mixed α/β ype.
α-helix: 1ny9.pdb (MW 10.60 kD, α-helix 73.4%, 3
10
-helix 3.2%, β-s and 0.0%,
u n 10.6%, coil 12.8%)
β-shee : 1 75.pdb (MW 7.85 kD, α-helix 0.0%, 3
10
-helix 0.0%, β-s and 23.5%, u n
26.5%, coil 50.0%)
α/β ype: 4ake.pdb (MW 23.59 kD, α-helix 43.5%, 3
10
-helix 5.1%, β-s and 15.4%,
u n 20.6%, coil 15.4%)
A icle No~e00175
3h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
In p epa a ion o ou o e all s udy, we o iginally chose p o eins iden ical o
ep esen a i es o hese domain old classes s udied in a ecen publica ion [16]
ha had as i s aim he in es iga ion o domain old space. The p o eins chosen
we e 1ny9, 1 75, and 1di .pdb (MW 16.31 kD, α-helix 32.2%, 3
10
-helix 6.0%,
β-s and 30.9%, u n 13.4%, coil 17.4%). Since hese p o eins a e ela i ely small
(sequence leng h, MW), a ou h p o ein, 4ake, was added o he s udy. Like 1di ,
4ake is a α/β ype, bu much la ge . Fu he , o he p esen s udy, we decided o
exclude 1di in lieu o 4ake, al hough he 1di p o ein ea u es in an ex ension o
his wo k, which is al eady unde way. The eason o excluding his p o ein a his
s age has o do wi h he s uc u al pa ame e s chosen o he i s pa o his wo k
(see below).
In he case o 1ny9, his is an NMR s uc u e con aining 10 al e na i e solu ions o
he s uc u e. The bes consensus s uc u e, no. 8, was de e mined using he WHAT
IF p og am [17]. The selec ed s uc u e he ein simply called 1ny9 is, s ic ly
speaking, he s uc u e wi h PDB iden i y numbe 1ny9 ha ing chain iden i ie A
and is he 8 h s uc u e in he NMR ensemble. This was chosen on he g ounds ha
i was he closes o he a e age s uc u e in ha ensemble and he e o e deemed o
be mos ep esen a i e.
The sequence o each o hese p o eins was hen submi ed o he P edic P o ein
se e [18]. P edic ed s uc u al pa ame e s we e ex ac ed om he ou pu iles
om his se e ( his was done manually bu an au oma ed e sion is unde
de elopmen ). The s uc u al pa ame e s included in ou schemes we e:
●SeqNo - Sequence numbe o a he posi ion unde conside a ion
●AA - iden i y o he amino acid esidue a he posi ion
●OHEL - Obse ed seconda y s uc u e
●PHEL - P edic ed seconda y s uc u e
●OACC - Obse ed sol en accessibili y
●PACC - P edic ed sol en accessibili y
The abo e a e all ob ained di ec ly as ou pu s om P edic P o ein. OHEL and
PHEL employ he HST e minology [19] whe eby H s ands o α-helix, S o
β-s and and T o u n. In p ac ice we employed he augmen ed se de ined wi hin
he WHAT IF p og am whe eby H, S, and T a e de ined as be o e [19] and
impo an seconda y s uc u e ea u es no included in HST a e added: 3 o 3
10
-
helix and C o coil. The impo ance o dis inguishing C om T is ha he o me
has he cha ac e o “ andom coil”while u ns con o m o s ingen limi s on
backbone o sion angles. N.B. in much o he bioin o ma ics li e a u e E
(“ex ended”) is used ins ead o S. Fo his wo k, a py hon sc ip was w i en o
A icle No~e00175
4h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
ex ac hese da a bu he e is e e y likelihood ha his will be ully au oma ed in
he nea u u e.
While he P edic P o ein p og am always e u ns a PHEL and PACC alue, i does
no always gi e he co esponding OHEL and OACC alues. In such cases, he
co esponding obse ed alues can be de e mined using WHAT IF, o o he
p o ein modelling so wa e. In ou case, we we e only in e es ed in OHEL and
OACC in aining examples. When a p o ein wi h unknown c ys al s uc u e is
s udied he da a will by de ini ion no be a ailable. I is o his eason ha PHEL
and PACC a e men ioned, since he e may be many easons why esea che s migh
wan o s udy p o eins ha lack a c ys al s uc u e (mu an s o example).
Finally, we included s uc u e in o ma ion no gene a ed by P edic P o ein:
AAH −Amino-acid hyd opa hy − alues o each o he 20 amino acid esidue
ypes a e ob ained om a da abase [20] ha p o ides he cu en ly mos eliable se
o da a.
The hyd opa hy da a a e an impo an ea u e o he o e all co pus o p o ein
s uc u al in o ma ion, bu he e is a p oblem in ha alues o ce ain pai s o
amino acid ypes, o example Asp and Glu, and A g and Lys, a e e y close o one
ano he , which a he masks he dis inc ly di e en s uc u al p opensi ies ha
hese esidue ypes ha e (see Table 1).
2.2. Soni ica ion o p o ein sequence and p ope y da a
The inpu da a o he h ee p o eins selec ed o his s udy a e lis ed in
supplemen a y ables 1–3. The columns con ain bo h expe imen ally de i ed da a
and da a de i ed om p edic ion me hods. The da a we e con e ed in o a o ma
ha can be assimila ed by he musicalgo i hms Web-based so wa e [21].
2.3. Da a soni ica ion mapping
In o de o disce n da a pa e ns mos clea ly as melodic exp essions, he da a- o-
music mappings ocused p ima ily on pi ch wi h uni o m hy hmic du a ions, and
pi ch wi h some hy hmic a ie y. Fu u e in es iga ions wi h musical e inemen s
should include seconda y cha ac e is ics such as imb e, dynamics, and a icula-
ions.
The da a- o-music mapping o amino acid chains was based on a p opo iona e
expansion algo i hm ha ans o ms amino acid alues o i a musical ange. The
amino acid alues om supplemen a y able 1 we e d awn om co esponding
whole- esidue hyd ophobici y measu emen s (a e a e aging) wi hin he wa e o
oc anol ange ob ained om he blanco.biomol.uci.edu Web si e, and lis ed in
Table 1. The mapping esul s can be displayed as a musical scale ascending om
A icle No~e00175
5h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).

a oma ic o cha ged ypes wi h 17 disc e e pi ches assigned o he ange o 20
amino acids (see Fig. 1 and supplemen a y Audio 1). No e, in h ee cases amino
acids we e pai ed up o sha e a single pi ch: M and Y, G and H, and A and S. The e
is no pa icula a p io i a ionale o hese pai ings, bu he e a e some possible
Table 1. AA hyd opa hy scale wi h a e aging; adap ed om
o iginal da a [20] and modi ied o enhanced disc imina ion.
AA AA-Hyd opa hy A e ages
W -2.1
F -1.7
L -1.3
I -1.1
Y -0.7
M -0.7
V -0.5
C 0.0
P 0.1
T 0.3
S 0.5
A 0.5
Q 0.8
N 0.9
G 1.2
H 1.2
R 1.8
E 1.9
D 2.0
K 2.8
[(Fig._1)TD$FIG]
Fig. 1. Musical scale c ea ed om AA whole- esidue hyd ophobici y measu emen s (a e a e aging).
A icle No~e00175
6h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
biophysical explana ions o he M/Y and A/S pai s. The la e esidue ypes bo h
ep esen he “small side-chain”class o esidues and hey can indeed “ eplace”
each o he −examples o en ies in mul iple sequence alignmen s ha ead
some hing like “AAAAASSSASAAAAASSAAAA”a e e y common. The M/Y
case is a bi mo e sub le. M is simila o Y bu also o F and W. They a e all
membe s o he “bulky side-chain”g oup and M can in many ins ances eplace he
o he h ee. The case o G and H, hese also need some explaining since H is
hough o in e ms o i s highly pola and ∼50% (a physiological pH) cha ged side
chain while G has no side chain. On he scale used he e, hese wo lie e y close
oge he , bu ha is also he case wi h se e al o he hyd opa hy scales. Mos o he
expe imen al echniques used o de e mine hese scales ely in one way o ano he
on pa i ioning be ween wo phases. We ha e o accep he ac ha H dis ibu es
i sel ac oss hese wo phases in much he same way as does G. In he aqueous
phase he ela i e popula ion o he cha ged species would be de e mined by he
pH, while in he nonpola phase only he uncha ged species would be p esen .
A p opo iona e expansion algo i hm was used o dis ibu e p o ein alues wi hin a
designa ed pi ch ange as de ined by he use . The e is lexibili y in se ing he
des ina ion pi ch span o wide o na ow anges, so ha use s can shape he musical
esul s acco ding o lis ene p e e ences. In any se ing, he esul s will demons a e
some p edic abili y, o example, T p (W), wi h an a e age hyd opa hy alue o
−2.09, will always ancho he lowes end o he chosen pi ch ange; Lys (K), wi h
an a e age hyd opa hy alue o 2.80, will always map o he highes end o he
pi ch ange; and Se (S), wi h an a e age hyd opa hy alue o 0.46 will map o he
middle o he pi ch ange. In his s udy, he AA alues we e mapped o a pi ch span
o 55 musical no es: piano keys 25–80 (see Fig. 2 and supplemen a y Audio 2).
This se ing p o ided a ela i ely e en dis ibu ion o independen pi ches o 17
AA alues. The same mapping dis ibu ion can apply o a ange o musical
du a ions. In hese cases a se o musical du a ions om sho o long can be
associa ed wi h nume ical da a alues om low o high; howe e , no hy hmic
a ie y was added a his ime so ha he complexi ies o he 17-pi ch scale could
be p esen ed in a simpli ied o m. Hence, each pi ch in he melodic ou pu was
ep esen ed by a uni o m hy hmic alue.
[(Fig._2)TD$FIG]
Fig. 2. In oduc o y exce p om 1ny9 sco e ep esen ing AA hyd ophobici y alues.
A icle No~e00175
7h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
The da a- o-music mapping o OHEL alues o e ed he challenge o con e ing
shapes in o music. This app oach equi ed a p elimina y s ep ha ans o med
OHEL da a in o “shaped”nume ic sequences, which we e hen soni ied o sho
musical ges u es −called mo i es and sub-ph ases:
Coil (C) = 1.0, 0.5, 0.8, 0.5, 1.0, 1.5, 1.2
Tu n (T) = 2.3, 1.7, 1.7, 2.0, 2.3
3
10
-helix (3) = 3.0, 2.8, 3.0, 3.2
α-helix (H) = 2.0, 1.6, 1.8, 2.2, 2.4
β-s and (S) o Ex ended (E) = 6.0, 5.8
The nume ic ep esen a ion o each shape can be lexibly mapped wi h he
p opo iona e expansion algo i hm o a a ie y o musical anges ha will expand
o comp ess he shapes on a e ical pi ch space (o axis). In music we call his
augmen ing o diminishing he in e als, which a e he dis ances be ween pi ches.
The choice o musical ange, whe he by expansion o comp ession, can also al e
he gene al onali y o he esul s and he size o he hy hmic du a ions. Hence he
choice o musical ange should be ca e ully conside ed o mee he goals o he
analysis and lis ene p e e ences o encou age epea ed lis ening.
The OHEL p o ein da a o 1 75 consis o a sequence o shapes labeled C, H, T, S.
Each p o ein esidue shape is mi o ed by a musical mo i e o simila con ou . In
his case he soni ica ion o one da a poin (o shape) will ea u e a sho collec ion
o pi ches a he han a single pi ch as seen in he p e ious mapping o AA
hyd opa hy. Fo example, he T (o u n) migh be ep esen ed by he nea ly
conjunc musical mo i e: Ab-F-F-F#-Ab. This descending and ascending musical
ges u e has a U shape, which mi o s he geome y o he T s uc u e (dihed al
angles a anged so ha he chain e e ses). As ano he example, 3
10
, a epea ing
egula s uc u e, is ep esen ed musically by epea ing pi ches as i in oscilla ion,
such as: B-Bb-B-C. In his s udy each OHEL shape om 1 75 is ep esen ed by a
mo i ic pa e n o pi ches and wi h a ied hy hms ha adhe e o he same
p opo iona e mapping sys em (see Fig. 3 and supplemen a y Audio 3), whe e
la ge alues such as 6.0 gene a e long hy hmic du a ions and small alues such as
0.5 become sho hy hmic du a ions.
The OACC alues co espond o su ace and dep h ela ions o he in e io o he
p o ein s uc u e. The OACC da a poin s a e mapped o he ull ange o he piano
o cap u e he a ying su ace- o-dep h ela ionships. In his melodic soni ica ion,
pi ch is he only poin o ocus as an e ec i e musical cha ac e is ic o ack da a
su ace o dep h ela ions. In his example, dep h is ep esen ed by he deep
sounding low pi ch con en , and su ace a eas by high pi ch con en . The uni o m
hy hmic du a ions a a s eady empo, o pulse, help he lis ene ocus on essen ial
A icle No~e00175
8h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
cha ac e is ics o ascending and descending pi ches (see Fig. 4 and supplemen a y
Audio 4).
The OACC alues we e mapped by means o he na u al log algo i hm. The log
algo i hm was selec ed o add pi ch space o he low and mid ange alues, gi ing
hem mo e sepa a ion o cla i y and de ini ion han he al e na i e p opo iona e
mapping sys em wi h equal dis ibu ion. The log mapping sys em is ideal as i
s e ches he a ailable low pi ch space and na ows he a ailable span o high
pi ch space o accommoda e a la ge po ion o low ange da a poin s o he p o ein
[(Fig._3)TD$FIG]
Fig. 3. In oduc o y exce p om 1 75 sco e ep esen ing OHEL alues.
[(Fig._4)TD$FIG]
Fig. 4. In oduc o y exce p om 4ake sco e ep esen ing OACC alues.
A icle No~e00175
9h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Table 2. Pe cep ual soni ica ion su ey esul s. OHEL and con ols used eal ime isual cues, while AA and OACC soni ica ion exce p s used no eal ime
isual cues. Pa icipan s we e able o pe cei e simila i ies be ween isual ep esen a ions and audible melodies gene a ed by he same da a.
Pa icipan s Soni ica ion Pe cep ual Su ey
(Human Subjec s 1-38)
Con ol An icipa ing Posi i e
("A" o "D" Responses)
Con ol An icipa ing Nega i e
("B" o "C" Responses)
AA 1ny9
exce p
OACC 4ake
exce p
OHEL 1 75
comple e
AA 1ny9 in
de ail
Subjec #1 C B A A D N/A
Subjec #2 A B A A A N/A
Subjec #3 A B C B D C
Subjec #4 A B C C A C
Subjec #5 D C D A A A
Subjec #6 D B A D D D
Subjec #7 D B D D D A
Subjec #8 D A C A D D
Subjec #9 A B D A D D
Subjec #10 A D A A C C
Subjec #11 D B A A D A
Subjec #12 A B A D A D
Subjec #13 A B D D A A
Subjec #14 A C A D A A
Subjec #15 A C A A D A
Subjec #16 A C D D A A
Subjec #17 D A B D D D
Subjec #18 A C A D D C
Subjec #19 D B B C A D
Subjec #20 A B A B A D
Subjec #21 A B A A A A
Subjec #22 D B B B C C
Subjec #23 D B A D D A
(Con inued)
A icle No~e00175
16 h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).

Table 2.(Con inued)
Pa icipan s Soni ica ion Pe cep ual Su ey
(Human Subjec s 1-38)
Con ol An icipa ing Posi i e
("A" o "D" Responses)
Con ol An icipa ing Nega i e
("B" o "C" Responses)
AA 1ny9
exce p
OACC 4ake
exce p
OHEL 1 75
comple e
AA 1ny9 in
de ail
Subjec #24 A B C C C A
Subjec #25 A B C C D A
Subjec #26 A B A A D A
Subjec #27 A B A A A A
Subjec #28 D B A A A A
Subjec #29 C B A A A A
Subjec #30 A C A A D A
Subjec #31 A C A D C D
Subjec #32 A B D C A C
Subjec #33 A B A A A A
Subjec #34 A C A A A A
Subjec #35 D C A A D A
Subjec #36 D B A C C D
Subjec #37 C B B B D B
Subjec #38 A B A A D A
To al Pe cen ages o Response Types
A Simila 60.53% 5.26% 60.53% 47.37% 42.11% 55.56%
B No Simila 0.00% 68.42% 10.53% 10.53% 0.00% 2.78%
C Somewha Simila 7.89% 23.68% 13.16% 15.79% 13.16% 16.67%
D Mos ly Simila 31.58% 2.63% 15.79% 26.32% 44.74% 25.00%
A icle No~e00175
17 h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
4. Discussion
We ha e sugges ed se e al ways in which p o ein sequences can be “hea d” a he
han “ ead”. The amino acid sequence is ende ed as a s ing o musical no es wi h
sound quali ies ha e lec he p ope ies o hese esidues. The e a e se e al
di e en ways o encoding hese p ope ies and we ha e examined se e al o hese,
all wi h in e es ing and memo able, bu di e en ou comes. Ce ain o hese
(OACC/PACC and AA) possess he p ope y o ha ing a 1:1 a io be ween pi ches
and esidues, while OHEL/PHEL has been cons uc ed speci ically in o de o
cap u e he de ining s uc u al ea u e o each seconda y s uc u e ype (e.g. 2
pi ches o S/E, 3 pi ches o 3
10
, and 4 pi ches o α-helix).
By lis ening o a melodic shape o a gi en p o ein sequence, an imp ession o i s
s uc u e can be buil up. We p opose his as an excellen lea ning ool o hose
wishing be e o unde s and and disc imina e p o ein old s uc u es, and as an aid
o old iden i ica ion i sel . We in end o ex end his wo k owa ds a mo e comple e
domain old iden i ica ion sys em, so ha wi h p ac ice, and by compa ison wi h
he sounds co esponding o o he p o ein ypes, his can lead o a way o
ecognizing he 3D olds o di e en p o eins. Fu he , we a e awa e ha p o eins
a e lexible en i ies ha ypically swi ch be ween wo di e en s uc u es in he
cou se o exe cising hei unc ion [22]. I would be in e es ing i ou musical
app oach could help o iden i y he signa u es co esponding o he esidue
posi ions ha a e esponsible o his pa icula ly impo an p o ein unc ion. The
musical pa e ns a e complex, and his will no only be because p o ein olds a e
complex bu also due o he need o swi ch be ween di e en s uc u al s a es. The
p ima y sequence o p o eins has o ca e o mo e han olding and swi ching
be ween olds, bu o a whole ange o o he unc ions [22] including how he
p o ein a i es a i s des ina ion in (o ou side) he cell, which equi es he
a achmen o pos - ansla ional signals on i s su ace. These unc ions a e also
encoded gene ically [22]. Fo hese easons we expec he e o be many
de elopmen s in he a ea whe e p o eomics and soni ica ion o e lap.
We should men ion ha simila a en ion is being paid o soni ica ion o DNA
sequences and gene exp ession [23,24]. These also encapsula e s uc u al
in o ma ion bu in a somewha di e en way, he p o ein sequences a e mo e
closely ela ed o he e en s ha ac ually ake place a he pheno ypic le el in he
li ing cell. These p o ein sequences a e encoded in DNA bu he la e can also
encompass egions ha ha e o do wi h cell di e en ia ion and epigene ic con ol
[24] which a e ou side o and dis inc om he p o ein coding egions. Thus he e
is e e y eason o engage in soni ica ion s udies a he DNA le el also.
A icle No~e00175
18 h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Decla a ions
Au ho con ibu ion s a emen
Robe P. Bywa e , Jona han N. Middle on: Concei ed and designed he
expe imen s; Pe o med he expe imen s; Analyzed and in e p e ed he da a;
Con ibu ed eagen s, ma e ials, analysis ools o da a; W o e he pape .
Funding s a emen
This wo k was suppo ed by an Eas e n Washing on Uni e si y Facul y Resea ch
and C ea i e Wo ks Summe G an . This wo k was also suppo ed in pa by Tekes
−The Finnish unding agency o inno a ion (decision 40296/14).
Compe ing in e es s a emen
The au ho s decla e no con lic o in e es .
Addi ional in o ma ion
Supplemen a y con en ela ed o his a icle has been published online a 10.1016/
j.heliyon.2016.e00175.
Acknowledgemen s
The au ho s would like o acknowledge he suppo o Eas e n Washing on
Uni e si y and he Uni e si y o Tampe e o helping us mo e soni ica ion
esea ch o wa d in he con ex o biological sciences.
Re e ences
[1] G. Bellesia, A.I. Jewe , J.E. Shea, Sequence pe iodici y and seconda y
s uc u e p opensi y in model p o eins, P o ein Sci. 19 (2010) 141–154.
[2] S. Ho mölle , T.P. Zhou, T. Ohlson, Con o ma ions o amino acids in
p o eins, Ac a C ys . D 58 (2002) 768–776.
[3] R.P. Bywa e , D. Thomas, G. V iend, A sequence and s uc u al s udy o
ansmemb ane helices, J. Compu -Aided Mol. Des. 15 (2001) 533–552.
[4] R.P. Bywa e , V. Ve yazo , The p e e ed con o ma ion o dipep ides in he
con ex o biosyn hesis, Na u wissenscha en 100 (2013) 853–859.
[5] A.G.Mu zin,S.E.B enne ,T.Hubba d,e al.,SCOP:as uc u al
classi ica ion o p o eins da abase o he in es iga ion o sequences and
s uc u es, J. Mol. Biol. 247 (1995) 536–540.
A icle No~e00175
19 h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
[6] C.A. O engo, A.D. Michie, S. Jones, e al., CATH −a hie a chic
classi ica ion o p o ein domain s uc u es, S uc u e 15 (1997) 1093–1108.
[7] N. Koga, R. Ta sumi-Koga, G.H. Liu, T.B. Ac on, G.T. Mon elione, D.
Bake , e al., P inciples o designing ideal p o ein s uc u es, Na u e 491
(2012) 222–227.
[8] S. Ba ass, P. Vicke s, Soni ica ion design and aes he ics, The Soni ica ion
Handbook, Logos Ve lag, 2011, pp. 145–172.
[9] B.N. Walke , G. K ame , Mappings and me apho s in audi o y displays: an
expe imen assessmen , ACM T ans. Appl. Pe cep . 2 (4) (2005) 407–412.
[10] The Soni ica ion Handbook, In: T. He mann, A. Hun , J.G. Neuho (Eds.),
Logos Ve lag, 2011.
[11] F. Dombois, G. Eckel, Audi ica ion, The Soni ica ion Handbook, Logos
Ve lag, 2011.
[12] J. Dunn, M.A. Cla k, Li e music: The soni ica ion o p o eins, Leona do 32
(1) (1999) 25–32.
[13] R. Takahashi, J.H. Mille , Con e sion o amino-acid sequences in p o eins o
classical music: Sea ch o audi o y pa e ns, Genome Biol. 8 (2007) 405.
[14] M.A. Ga cia-Ruiz, J.R. Gui e ez-Pulido, An o e iew o audi o y display o
assis comp ehension o molecula in o ma ion, In e ac . Compu . 18 (2006)
853–868.
[15] A. Suppe , Sublime equencies: he cons uc ion o sublime lis ening
expe iences in he soni ica ion o scien i ic da a, Soc. S ud. Sci. 44 (1) (2014)
34–58.
[16] P. Mina y, M. Le i , P obing p o ein old space wi h a simpli ied model, J.
Mol. Biol. 375 (2008) 920–933.
[17] G. V iend, WHAT IF: a molecula modelling and d ug design p og am, J.
Mol. G aph. 8 (1990) 52–56.
[18] B. Ros , G. Yachda , J.F. Liu, The P edic P o ein se e , Nucl. Acids Res. 32
(2004) W321–W326.
[19] W. Kabsch, C. Sande , Dic iona y o p o ein seconda y s uc u e: pa e n
ecogni ion o hyd ogen-bonded and geome ical ea u es, Biopolyme s 22
(1983) 2577–2637.
[20] W.C. Wimley, S.H. Whi e, Memb ane pa i ioning: Dis inguishing bilaye
e ec s om he hyd ophobic e ec , Biochemis y 32 (1993) 6307–6312.
A icle No~e00175
20 h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
[21] J. Middle on, D. Dowd, Web-based algo i hmic composi ion om ex amu-
sical esou ces, Leona do 41 (2) (2008) 128–135.
[22] R.P. Bywa e , P o ein olding: a p oblem wi h mul iple solu ions, J. Biomol.
S uc . Dyn. 31 (4) (2013) 351–362.
[23] M.S. S aege, A sho ea ise conce ning a musical app oach o he
in e p e a ion o gene exp ession da a, Sci. Rep. 5 (2015) 15281.
[24] D. B ocks, Musical pa e ns o compa a i e epigenomics, Clin. Epigene ics 7
(2015) 94.
A icle No~e00175
21 h p://dx.doi.o g/10.1016/j.heliyon.2016.e00175
2405-8440/© 2016 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).