Download msb200922-sup

Predicting Metabolic Biomarkers of Human Inborn Errors of
Metabolism
Supplementary Information
Tomer Shlomi1†, Moran N. Cabili2†, Eytan Ruppin2, 3
1
Department of Computer Science, Technion – Israel Institute of Technology, Haifa
32000, Israel
2
School of Computer Science, Tel-Aviv University, Tel-Aviv 69978, Israel
3
School of Medicine, Tel-Aviv University, Tel-Aviv 69978, Israel
†
These authors contributed equally to this work
Table of Content
Supplementary Figure 1
An illustrative example of biomarker's extra-cellular concentration change
prediction………………………………………………………………………………..
Supplementary Figure 2
Distribution of biomarkers in different biofluids according to HMDB ………………..
Supplementary Figure 3
Illustration of the effect of Tyrosinemia type I, type III and Alkaptonuria on the
metabolism and transport of tyrosine…………………………………………………...
Supplementary Figure 4
Illustration of the effect of Methylmalonate semialdehyde dehydrogenase deficiency
on the metabolism and transport of its known biomarkers……………………………..
Supplementary Table 1
Full names of the metabolites presented in main-text figure 3…………………………
Supplementary Table 2
Full names of the metabolites presented in supplementary figure 3……………………
Supplementary Table 3
Full names of the metabolites presented in supplementary figure 4……………………
2
3
4
5
6
7
8
1
Supplementary Figure 1: An illustrative example of the computed flux distributions
underlying the predicted reduction in the extra-cellular concentration of biomarker M6.
Circular nodes represent metabolites, solid arrows represent reactions, and dashed lines
represent the network boundaries. The disease causing reaction is marked with a red
cross. A predicted steady-state flux distribution is colored orange. (a) A flux distribution
describing the maximal flux value of V6 in the healthy state. M5 is converted to M6 and
2
some of M6 is secreted from the cell. (b) A flux distribution describing the minimal flux
value of V6 in the healthy state. (c) A flux distribution describing the maximal flux value
of V6 in the disease state. (d) A flux distribution describing the minimal flux value of V6
in the disease state.
Supplementary Figure 2: Number of predicted biomarkers in disorders that can be
found in different biofluids according to HMDB (data is available for 137 out of 223
biomarkers predicted in total). According to HMDB the biomarkers are classified to
different metabolic sub-classes that are shown in stack. Metabolites that were classified
to other metabolic sub-classes or those for which no classification was available are
classified as 'other'.
3
Supplementary Figure 3: A sub-network that illustrates the effect of Tyrosinemia type
I, type III and Alkaptonuria (OMIM 276700, 276710, 203500, respectively) on the
metabolism and transport of tyrosine (tyr). Circular nodes represent metabolites and
edges represent biochemical reactions. For simplicity, only abbreviations of metabolite
names and enzyme E.C. (Enzyme Commission) numbers are specified (explicit names
are given in Supp. Table 2). Metabolites marked in green participate in other reactions
that are not presented here, for simplicity. Tyrosinemia type III, Alkaptonuria, and
Tyrosinemia type I are caused by dysfunctional HPD, HGD, and FAH, respectively.
As shown, tyrosine can be degraded via 5 different pathways. Interestingly though, our
method correctly predicts that the extra-cellular concentration of tyrosine would increase
in all aforementioned disorders, though a visual inspection of the network topology may
incorrectly suggest otherwise (i.e. that the alternative pathways would fully compensate
for the dysfunctional enzymes and maintain the same uptake rate of tyrosine in a disease
case).
4
Supplementary Figure 4: A sub-network that illustrates the effect of Methylmalonate
semialdehyde dehydrogenase deficiency (OMIM 603178) on the metabolism and
transport of valine (val), beta-alanine (ala_B), 3-hydroxypropionic acid (3hpp), 3aminoisobutyric acid (3aib) and 3-hydroxyisobutyric acid (3hmp). Circular nodes
represent metabolites, full edges represent biochemical reactions and a dashed edge
represents a biochemical pathway. Metabolites marked in green participate in other
reactions that are not presented here for simplicity. Only abbreviations of metabolite
names and enzyme E.C. (Enzyme Commission) numbers are specified (explicit names
are given in Supp. Table 3). Methylmalonate semialdehyde dehydrogenase deficiency is
caused by a dysfunctional ALDH6A1. In this case, our method fails to predict the
elevation of 3hpp and 3hmp as documented in OMIM. An inspection of network
topology reveals that these false predictions result from missing membrane transporters
for these products, which prevent their secretion and hence indirectly limit the possible
uptake of their upstream substrate, valine – leading also to the false prediction of an
elevated extra-cellular concentration of valine (val). These newly hypothesized
5
transporters are marked as red edges in the figure (notably though, our method correctly
predicts the elevated extra-cellular concentration of 3aib in this case).
Short name
Full name
2kmb
2-keto-4-methylthiobutyrate
dkmpp
2,3-diketo-5-methylthio-1-phosphopentane
5mdru1p
5-Methylthio-5-deoxy-D-ribulose1-phosphate
5mdr1p
5-Methylthio-5-deoxy-D-ribose1-phosphate
5mta
5-Methylthioadenosine
ametam
S-Adenosylmethioninamine
amet
S-Adenosyl-L-methionine
ahcys
S-Adenosyl-L-homocyteine
hcys
L-Homocysteine
ser
L-Serine
cyst
L-Cystathionine
cys
L-Cysteine
met
L-Methionine
Supplementary Table 1: Full names of the metabolites presented in main-text figure 3.
6
Short name
Full name
tyr
L-Tyrosine
34hpp
3-(4-Hydroxyphenyl)pyruvate
akg
2-Oxoglutarate
glu
L-Glutamate
hgentis
Homogentisate
4mlacac
4-Maleylacetoacetate
4fumacac
4-Fumarylacetoacetate
acac
Acetoacetate
fum
Fumarate
tym
Tyramine
3ityr
3-Iodo-L-tyrosine
i
Iodide
iod
Iodine
34dhphe
3,4-Dihydroxy-L-phenylalanine
thbpt
Tetrahydrobiopterin
thbpt4acam
Tetrahydrobiopterin-4a-carbinolamine
phe
L-Phenylalanine
Supplementary Table 2: Full names of the metabolites presented in supplementary
figure 3.
7
Short name
Full name
val
L-Valine
ala_B
beta-Alanine
3hpp
3-Hydroxypropanoate , (syn: 3-Hydroxypropionic acid)
3hpcoa
3-Hydroxypropionyl-CoA
msa
Malonate semialdehyde
3hmp
3-Hydroxy-2-methylpropanoate, (syn: 3-Hydroxyisobutyric acid )
2mop
2-Methyl-3-oxopropanoate, (syn: Methylmalonate semialdehyde)
accoa
Acetyl-CoA
ppcoa
Propanoyl-CoA
2maacoa
2-Methyl-3-acetoacetyl-CoA
3aib
L-3-Amino-isobutanoate , (syn:3-aminoisobutyric acid)
Supplementary Table 3: Full names of the metabolites presented in supplementary
figure 4.
8