Volume 9 Supplement 1

## Selected articles from the IEEE International Conference on Bioinformatics and Biomedicine 2010

# Bayesian non-negative factor analysis for reconstructing transcription factor mediated regulatory networks

- Jia Meng
^{1}, - Jianqiu (Michelle) Zhang
^{1}, - Yidong Chen
^{2, 3}and - Yufei Huang
^{1, 2, 3}Email author

**9(Suppl 1)**:S9

**DOI: **10.1186/1477-5956-9-S1-S9

© Meng et al; licensee BioMed Central Ltd. 2011

**Published: **14 October 2011

## Abstract

### Background

Transcriptional regulation by transcription factor (TF) controls the time and abundance of mRNA transcription. Due to the limitation of current proteomics technologies, large scale measurements of protein level activities of TFs is usually infeasible, making computational reconstruction of transcriptional regulatory network a difficult task.

### Results

We proposed here a novel Bayesian non-negative factor model for TF mediated regulatory networks. Particularly, the non-negative TF activities and sample clustering effect are modeled as the factors from a Dirichlet process mixture of rectified Gaussian distributions, and the sparse regulatory coefficients are modeled as the loadings from a sparse distribution that constrains its sparsity using knowledge from database; meantime, a Gibbs sampling solution was developed to infer the underlying network structure and the unknown TF activities simultaneously. The developed approach has been applied to simulated system and breast cancer gene expression data. Result shows that, the proposed method was able to systematically uncover TF mediated transcriptional regulatory network structure, the regulatory coefficients, the TF protein level activities and the sample clustering effect. The regulation target prediction result is highly coordinated with the prior knowledge, and sample clustering result shows superior performance over previous molecular based clustering method.

### Conclusions

The results demonstrated the validity and effectiveness of the proposed approach in reconstructing transcriptional networks mediated by TFs through simulated systems and real data.

## Background

Transcription factor is one major gene regulator that governs the response of cells to changing endogenous or exogenous conditions [1]. Understanding how transcriptional regulatory networks (TRNs) induce cellular states and eventually define the phenotypes represents a major challenge facing systems biologists. So far, numerous models have been proposed to infer the transcriptional regulations by TFs including, ordinary differential equations, (probabilistic) Boolean networks, Bayesian networks, and information theory and association models, etc [2]. Ideally, the TF protein level activities are needed for exact modeling; however, due to low protein coverage and poor quantification accuracy of high throughput proteomics technologies such as protein array and liquid chromatography-mass spectrometry (LC-MS), the measurements of TF protein activities are currently hardly available. As a compromise, most of the aforementioned models conveniently yet inappropriately assume the TF’s mRNA expression as its protein activity. Given the fact that gene mRNA expression and its protein abundance are poorly correlated [3, 4], these models cannot accurately model the transcriptional *cis*-regulation or reveal at the best TF *trans*-regulation.

In contrast, works based on factor models [5–10] point to a natural and promising direction for modeling the TF mediated regulations, where the microarray gene expression is modeled as a linear combination of unknown TF activities, and the loading matrix in this model indicates the strength and the type (up- or down- regulation) of regulation. However, due to distinct features of TF regulation, conventional FA model is not readily applicable. First, due to various reasons (normal and disease, cancer grade, subtypes, etc), the samples are usually not independent with each other but show some clustering effect; while in the existing FA models, factors are typically assumed independent, which, although true in many applications, is not a realistic assumption for TF medicated regulation. Secondly, since a TF only regulates a small subset of genes, the loading matrix should be sparse. While with constructions of TF regulation databases, such as TRANSFAC [11], the knowledge of TF regulated genes becomes increasingly available, and should be included in the model so as to boost signal-to-noise and improve performance [12]. The inclusion of prior information and sparsity constraint naturally call for a Bayesian solution. As an added advantage, having this prior knowledge actually resolves the factor order ambiguity of the conventional factor analysis. Thirdly, as suggested in [13–15], the non-negative assumption on TF activities be imposed.

In a response to these requirements of modeling TF mediated regulatory networks, we propose here a novel Bayesian non-negative factor model (BNFM). Different from conventional factor analysis models, BNFM consists of a sparse loading matrix and a set of correlated non-negative factors. The sparsity of the loading matrix is constrained by a sparse prior [16] that directly reflects our existing knowledge of TF regulation. That is if a gene is known to be regulated by a TF, then the prior probability that this regulation exists is high, and otherwise, very low due to the generic sparse nature of TF regulation (A TF only regulates a small number of genes in the whole genome). Because of clustering effect on the data samples, the factors in this BNFM model are considered to be correlated and modeled by a Dirichlet process mixture (DPM) prior [17]. DPM imposes a natural non-parametric clustering effect [18] among samples of the same TF and can automatically determine the optimal number of clusters. Moreover, since the activities of TFs are non-negative, they are assumed to follow a (non-negative) rectified Gaussian distribution [19]. Due to the complex nonlinear structure of the BNFM, the estimation of the model becomes analytically infeasible and highly complicated numerically. A Gibbs sampling solution is developed to infer all the relevant unknown variables.

## Method

### Bayesian non-negative factor model

*n*-th microarray mRNA expression profile of

*G*genes under a specific context. In practice, microarray data

**y**

_{ n }register the log2 scaled (fold change of) gene expression levels under the context of interest relative to a background often obtained as the average expression levels among a variety of contexts, such as different cell lines and tumors [20, 21]. We assume that the expression level

**y**

_{ n }is due to the linear combination of scaled TF absolute protein activites and modeled by the following factor model

where,

**x**

_{ n }- the

*n*-th sample vector of the scaled activities of

*L*TFs of interest. Particulary, the non-negativity of

**x**

_{ n }is modeled by applying the component-wise rectification (or cut) function

*cut*to a vector pseudo factors

**s**

_{ n }, such that the

*l*-th element of

**x**

_{ n }is expressed as

**s**

_{ n }are modeled by a Dirichlet Process Mixture (DPM) of the Gaussian distributions as

*µ*

_{ l }

_{,}

_{ n }and variance ,

*DP*denotes the Dirichlet process, and

*NIG*is short for the conjugate Normal-Inverse-Gamma (NIG) distribution. This DPM model implies a clustering effect on

**s**

_{ n }such that

*γ*

_{ n }∈ ℤ represents the cluster label of the

*n*-th sample and is governed by a discrete GEM distribution [17], which defines the stick breaking process with parameter

*α;*this implies that the elements of

**s**

_{ n }are correlated. Based on (2) and (3), we have

where,
denotes the rectified Gaussian distribution [19]. Since
and *γ*
_{
n
} are still defined in (4) by the DP, **x**
_{
n
} is hence modeled by the DPM of the rectified Gaussian distributions and the elements of **x**
_{
n
} are accordingly correlated. In contrast to the conventional mixture model, the DPM model enables the number of clusters to be learnt adaptively from the data instead of being predefined.

**A**- the

*G*×

*L*loading matrix, whose element

*a*

_{ g }

_{,}

_{ l }represents the regulatory coefficient of the

*g*-th gene by the

*l*-th TF. Since a TF is known to regulate only small set of genes, A should be sparse. In our model, the elements of A are assumed to be independent and with the

*a priori*distribution [16]

*π*

_{ g }

_{,}

_{ l }is the

*a priori*probability of

*a*

_{ g }

_{,}

_{ l }to be nonzero. For instance, if a TF regulates a total of 500 genes among the 20000 genes in the human genome, then

*π*

_{ g }

_{,}

_{ l }is equal to

In most cases, *π*
_{
g
}
_{,}
_{
l
} are likely to be smaller than 10%. In practice, databases such as TRANSFAC [11] and DBD [22] provide information of experimentally validated or predicted target genes of TFs, and this knowledge can be incorporated in the model by setting, for instance, *π*
_{
g
}
_{,}
_{
l
} = 0*.* 9, if TF *l* is known to regulate gene *g;* or otherwise *π*
_{
g
}
_{,}
_{
l
} = 0*.* 025. The variable
defines how much the target genes are *loaded* on the corresponding TF and with prior distribution
.

**c**- a vector of constant, which can be considered as the constant term retained when linearizing the general relationship **y**
_{
n
} = *f*(**x**
_{
n
}) as **y**
_{
n
} = **Ax**
_{
n
} + c. It may also be interpreted as static response of gene transcriptional expressions.

**e**
_{
n
}- the *G* × 1 white Gaussian noise vector characterized by the covariance matrix
and with prior distribution
.

### Equivalent model for centralized observations

**ŷ**

_{ n }, where,

**ŷ**

_{ n }=

**y**

_{ n }– µ

_{ y }and . Traditionally, since the models typically assume zero mean for the factors, the equivalent model for centralized observations remains the same except that the constant term is eliminated, i.e., if

**y**

_{ n }=

**Ax**

_{ n }+

**c**+

**e**

_{ n }, then, for the centralized data

**ŷ**

_{ n },

_{ y }can be viewed as an ML estimator of the constant term c [23]. For BNFM, however, since the factor mean is no longer zero, the equivalent model for BNFM no longer remains the same as above mentioned traditional model, but instead,

_{ x }= [

*µ*

_{ x }

_{1},

*µ*

_{ x }

_{2}, …,

*µ*

_{ x }

_{L}]

^{⊺}can be approximated with the mean of prior distribution (4)(5), which can be calculated numerically. We can also see that the corresponding centralized factors are a shifted version of the original factors, and different samples shift the same amount, so sample clustering effect is still retained. On the other hand, the removed term from data centralization is no longer an estimator of the constant term

**c**, but,

The goal is to obtain the posterior distributions and hence the estimates of **A**, **x**
_{
n
}∀*n*, *γ*
_{
n
}∀*n*, given **y**
_{
n
}∀*n* and *π*
_{
g
}
_{,}
_{
l
} ∀*g*, *l*, which is the TF binding prior information extracted from existing database. For convenience, we let **Θ** denote all the known and unknown variables.

## Gibbs sampling solution

The proposed BNFM model is high dimensional and analytically intractable, so a Gibbs sampling solution is proposed. Gibbs sampling devises a Markov Chain Monte Carlo scheme to generate random samples of the unknowns from the desired but intractable posterior distributions and then approximate the (marginal) posterior distributions with these samples. The key of Gibbs sampling is to derive the conditional posterior distributions and then draw samples from them iteratively until convergence is reached. The proposed Gibbs sampler can be summarized as follows:

### Gibbs sampling for BNFA

*t*-th iteration:

- 1.
Sample from ;

- 2.
Sample from ;

- 3.
Sample from

*p*(*a*_{ g }_{,}_{ l }|**Θ**_{–a }_{ g,l }); - 4.
for

*n*= 1 to*N*

Sample from ;

Sample from ;

Sample
from *p*(**s**
_{
n
}|**Θ**
_{–s
}_{
n
});

Note that
are marginalized and therefore does not need to be sampled. The algorithm iterates until the convergence of samples, which can be assessed by the scheme described in [24], [chap. 11.6]. The samples after convergence will be collected to approximate the marginal posterior distributions and the estimates of the unknowns. Since µ
_{
x
} can be approximated and calculated numerically, the factor **x**
_{
n
} can be recovered from the centralized factor
with (9). The required conditional distributions of the above proposed Gibbs sampling solution are detailed in the next.

### Conditional distributions of the proposed Gibbs sampling solution

For simplicity, we let **x**
_{
n
} and **y**
_{
n
} denote the centralized factors and data in this section.

**E**=

**Y**–

**AX**and

**e**

_{ g }= [

*e*

_{ g }

_{,1},

*⋯*,

*e*

_{ g }

_{,}

_{ N }]

^{T}, then,

**ŷ**

_{ g }

_{,}

_{ l }=

**x**

_{ l }

*a*

_{ g }

_{,}

_{ l }+

**e**

_{ g },

**x**

_{ l }= [

*x*

_{ l }

_{,1},

*⋯*,

*x*

_{ l }

_{,}

_{ n }]

^{T}and

**e**

_{ g }= [

*e*

_{ g }

_{,1},

*⋯*,

*e*

_{ g }

_{,}

_{ N }]

^{T}

*.*The posterior distribution of

*a*

_{ g }

_{,}

_{ l },

*Z*

_{0}is a normalizing constant, is the posterior probability of

*a*

_{ g }

_{,}

_{ l }≠ 0 and

*BF*

_{01}is the Bayes factor of model

*a*

_{ g }

_{,}

_{ l }= 0 versus model

*a*

_{ g }

_{,}

_{ l }≠ 0

*a*

_{ g }

_{,}

_{ l }≠ 0 and defined by

where,
and
, and *π*
_{
g
}
_{,}
_{
l
} is the prior knowledge of the probability of *a*
_{
g
}
_{,}
_{
l
} to be non-zero. When *π*
_{
g
}
_{,}
_{
l
} = 0*.* 5, i.e, a noninformative prior on sparsity is assumed,
depends only on *BF*
_{01}, and
when *BF*
_{01} > 1. Since model selection based *BF*
_{01} favors *a*
_{
g
}
_{,}
_{
l
} = 0, it suggests that this Bayesian solution favors sparse model even when *π*
_{
g
}
_{,}
_{
l
} = 0*.* 5.

*γ*

_{ n }does not depend on

**x**

_{ n }in the distribution. It is intended that samples of

*γ*

_{ n }from this distribution are not affected by the immediate sample of

**x**

_{ n }, thus achieving faster convergence of the sample Markov chains. To derive this distribution, first let

**ŷ**

_{ l }

_{,}

_{ n }=

**a**

_{ l }

*x*

_{ l }

_{,}

_{ n }+

**e**

_{ n }with

**a**

_{ l }being the

*l*-th column of

**A**and hence . Then,

*s*

_{ l }that also belong to cluster

*k*,

*N*

_{–l,k }is size of , and

Noted that, for a new cluster,
and *N*
_{
–l
}
_{,}
_{
k
} = 0, and
can be derived from *g*
_{
k
} for
similarly.

*µ*

_{ x },+∞), and,

*x*

_{ l },

*n*, the conditional distribution of

*s*

_{ l }

_{,}

_{ n }does not depend on

**y**

_{ 1:N }; As the predictive density

*p*(

*s*

_{ l }

_{,}

_{ n }|

**s**

_{–}

_{ l }

_{,}

_{ n },

*γ*

_{ l }) is shown to be a Student-t distribution, which can be conveniently approximated as a normal distribution when

*N*

_{ –l }

_{,}

_{ k }is large:

where, *π*
_{
s
}_{
l,n
} = 1 – sgn (*x*
_{
l
}
_{,}
_{
n
} + *µ*
_{
x
})

**E**=

**Y**–

**AX**, and we have, , where

**e**

_{ g }= [

*e*

_{ g }

_{,1},

*e*

_{ g }

_{,2},

*…*,

*e*

_{ g }

_{,}

_{ N }]

^{⊺}Given the conjugate Inverse-Gamma prior, we have

where,
and
, and *N*
_{
a
}
_{,}
_{
l
} is the size of
.

## Results

### Test on simulated system

The proposed BNFM model was first tested on a simulated system, in which the microarray data consists of the expression profiles of 150 genes with 40 samples. The samples form 5 clusters and the 150 genes were assumed to be regulated by 10 TFs. The sparsity of loading matrix was set at 10%, which means that on average each gene is regulated by 1 TFs, and each TF regulates 15 genes. To simulate a practical imperfect database, the precision and recall of the prior knowledge were both set equal to 0.9 each, i,e., 90% of the database recorded regulations indeed happened in this specific data set (10% of the database recorded regulations may be context-specific and didn’t happen in the data); and 90% of the true regulations was recorded in the database (10% of true regulations are not in the database). This setting indicates that the recorded prior regulations may not exist in the experiment, and the unknown regulations could exist. Since this is a relatively large data set involving sampling of many variables, instead of examining convergence based on [24], [chap. 11.6], we adopted a more practical strategy by running a single MCMC chain for 10000 iterations with a burn-in period of 2000 iterations [25].

*F*metric [26] that combines the BCubed precision and recall [27] was implemented as suggested in [28]. More specifically, let

*L*(

*e*) and

*C*(

*e*) be the category and the cluster of an item

*e*. Then, the correctness of the relation between

*e*and

*e*

^{ ′ }is defined by

*F*metrics:

The *F* metrics satisfy all the 4 formal constraints defined in [28] including cluster homogeneity, cluster completeness, rag bag, and cluster size vs. quantity. We adopt the *F* metrics to evaluate the clustering result in the following tests. Similarly, a Van Rijsbergen’s *F* metric that combines the target prediction precision and recall is used to measure the target prediction result. Since our model can avoid sign ambiguity problem, the loading and factor estimations were evaluated using its Pearson’s correlation with their true values.

### Test on breast cancer data

^{+}breast cancer. All samples came with gene microarray expression, ER status and survival time information. For the settings of the algorithm, we first manually selected a total of 15 TFs that are reported to be relevant to breast cancer (Table 1) and then retrieved a total of 199 regulated target genes (Table 2) by these TFs from TRANSFAC database [11] (Release 2009.4). We assume that TRANSFAC record has a 90% precision and 90% recall, suggesting that the known regulations may be context-specific and unknown regulations could exist. From the precision and the recall, the prior probability of the loading matrix can be determined. Based on these settings, the proposed approach was applied to the breast cancer data set to infer the underlying regulatory networks and TF activities. The posterior distribution of the loading matrix (Fig.5) gives insight into the sparsity of inferred TF mediated regulation. It can be seen that the posterior probability of regulations fall into 2 distinct groups, i.e., one group has very small posterior probabilities, which correspond to regulations that do not exist; while the other group have larger posterior probabilities, which correspond the regulations that are likely to exist. Fig. 6 depicts possible regulations and their posterior probabilities (rounded by 0.1) in a network, demonstrating the capability of the proposed approaches to identify possible TF regulated target genes.

List of tested 15 TFs and aliases

Name | Target | Aliases | |
---|---|---|---|

1 | EBP- | 21 | BPc; C/EBP; C/EBP alpha; C/EBPalpha; CBP; |

2 | ETS-1 | 14 | c-Ets-1; c-Ets-1 54; c-Ets-1A; Ets1; p54; p54c-Ets-1. |

3 | FOS | 23 | c-Fos; FBJ osteosarcoma oncogene; p55(c-fos). |

4 | MYC | 11 | c-Myc; MYC; v-myc myelocytomatosis viral oncogene homolog (avian). |

5 | CREB | 22 | ATF-47; CREB; CREB-341; CREB-A; CREB-isoform1; CREB1; |

6 | ATF-2 | 16 | activating transcription factor 2; ATF2; CRE-BP1; CREB2; CREBP1; |

7 | EGR-1 | 24 | AT225; early growth response protein 1; |

8 | EBP- | 23 | AGP/EBP; ANF-2; C/EBP beta; C/EBP-beta; C/EBPbeta; CEBPB; |

9 | NF- | 28 | NFkappaB; Nuclear Factor kappa B. |

10 | P53 | 20 | ASp53; LFS1; NSp53; p53; p53as; RSp53; tp53; TRP53; |

11 | ATF-1 | 14 | activating transcription factor 1; ATF1; EWS-ATF1; FUS/ATF-1; |

12 | STAT-3 | 12 | acute-phase response factor; APRF; |

13 | STAT-1 | 19 | signal transducer and activator of transcription 1. |

14 | AP-2 | 16 | activating enhancer binding protein 2 alpha; activator protein-2; |

15 | CREB-1 | 19 | ATF-47; CREB; CREB1; cyclic AMP response element-binding protein; |

List of tested 199 genes

Symbol | Symbol | Symbol | Symbol | ||||
---|---|---|---|---|---|---|---|

1 | CXCR4 | 51 | CD82 | 101 | PENK | 151 | CDKN1A |

2 | CAT | 52 | HLA-DRA | 102 | PIM1 | 152 | PTTG1 |

3 | FOS | 53 | VIP | 103 | COL1A2 | 153 | MITF |

4 | MT2A | 54 | INS | 104 | IL2RB | 154 | HBB |

5 | PSMB9 | 55 | PTGS2 | 105 | ZNF268 | 155 | CSF1 |

6 | DBH | 56 | APOA2 | 106 | GSN | 156 | TIMP1 |

7 | SERPINC1 | 57 | FGFR2 | 107 | TNFRSF10C | 157 | F9 |

8 | CHEK1 | 58 | CCND1 | 108 | CXCL3 | 158 | VHL |

9 | SCN3B | 59 | CASP1 | 109 | CSNK2B | 159 | CD1A |

10 | F7 | 60 | HBB | 110 | TRA@ | 160 | SFN |

11 | ITGAX | 61 | COL2A1 | 111 | HLA-DPB1 | 161 | SOAT1 |

12 | EIF4E | 62 | MDM2 | 112 | TRA@ | 162 | FCGR1A |

13 | TGFB2 | 63 | RB1 | 113 | TP53 | 163 | FAS |

14 | CDC25A | 64 | NDRG1 | 114 | SOX9 | 164 | HBG1 |

15 | IL3 | 65 | BRCA1 | 115 | ALOX5AP | 165 | WARS |

16 | SERPINE1 | 66 | BAX | 116 | TOP1 | 166 | KIR3DL1 |

17 | IL10 | 67 | ATF2 | 117 | NFKB1 | 167 | CD8A |

18 | F3 | 68 | FN1 | 118 | IL2 | 168 | IL6 |

19 | IL2RA | 69 | BCL2L1 | 119 | SLC9A3 | 169 | TWIST1 |

20 | BDNF | 70 | CCR5 | 120 | CYP3A4 | 170 | CXCL1 |

21 | WEE1 | 71 | TF | 121 | CRH | 171 | IFNB1 |

22 | CYP11A1 | 72 | TFRC | 122 | CIITA | 172 | PTK2 |

23 | NR4A2 | 73 | HD | 123 | RFWD2 | 173 | SPP1 |

24 | VHL | 74 | CXCL1 | 124 | LOR | 174 | CSF1 |

25 | TRH | 75 | CSNK1A1 | 125 | REN | 175 | TP73 |

26 | SOD2 | 76 | NR3C1 | 126 | YBX1 | 176 | CD53 |

27 | CSF2RA | 77 | SPINK1 | 127 | ATF3 | 177 | NAB2 |

28 | MUC1 | 78 | EGR1 | 128 | TEAD1 | 178 | PTTG1 |

29 | MEFV | 79 | EDN1 | 129 | CDK4 | 179 | IL1B |

30 | GNAI2 | 80 | TFAP2A | 130 | APAF1 | 180 | APOB |

31 | DRD1 | 81 | CFTR | 131 | CYP19A1 | 181 | IL8 |

32 | ADRB2 | 82 | MYC | 132 | ACE | 182 | TAF7 |

33 | GCLC | 83 | FMR1 | 133 | KRT16 | 183 | PTP4A1 |

34 | OPRM1 | 84 | F8 | 134 | NOS2A | 184 | HSD17B8 |

35 | IFNG | 85 | TSC22D3 | 135 | FXR2 | 185 | ABCB1 |

36 | BCL2A1 | 86 | FGF2 | 136 | IRF1 | 186 | PBK |

37 | CCL5 | 87 | LOR | 137 | CGA | 187 | TACR1 |

38 | ICAM1 | 88 | PTHLH | 138 | KRT14 | 188 | MAOB |

39 | PSENEN | 89 | S100A9 | 139 | ABCA2 | 189 | RPL10 |

40 | IER2 | 90 | GADD45A | 140 | FGA | 190 | IVL |

41 | SOD1 | 91 | EXO1 | 141 | TALDO1 | 191 | ERBB2 |

42 | GNRHR | 92 | PLAU | 142 | CSF1 | 192 | CCL2 |

43 | LTA | 93 | PTH | 143 | SFTPD | 193 | BBC3 |

44 | TERT | 94 | CDK4 | 144 | CRP | 194 | TP63 |

45 | TNFAIP6 | 95 | PPARG | 145 | TPT1 | 195 | RFWD2 |

46 | ODC1 | 96 | POLB | 146 | SLC9A2 | 196 | FGFR4 |

47 | LTF | 97 | ID1 | 147 | CYP2A13 | 197 | NAT1 |

48 | PRLR | 98 | MT2A | 148 | DDX18 | 198 | SELE |

49 | TNF | 99 | SST | 149 | CCNA2 | 199 | FASLG |

50 | MMP1 | 100 | KRT14 | 150 | IL6ST |

*p*= 0

*.*05) as shown in Fig.9. Previous studies based on expression levels [33–36] identified 5 major subtypes (luminal A, luminal B, basal, ERBB2 overexpressing, and normal-like). We compared the pair-wise survival difference between our clustering (3 clusters) and previous result (5 clusters). It shows the superior performance of our method (Table 3) over the previous computation based method (Table 4).

Survival test of clustering results

Cluster 1 | Cluster 2 | Cluster 3 | |
---|---|---|---|

Cluster 1 | N/A | 0.04 | 0.16 |

Cluster 2 | 0.04 | N/A | 0.93 |

Cluster 3 | 0.16 | 0.93 | N/A |

Survival test of previous results

luminal A | luminal B | Basal-like | HER2+/ER- | normal-like | |
---|---|---|---|---|---|

luminal A | N/A | 0.75 | 0.76 | 0.42 | 0.83 |

luminal B | 0.75 | N/A | 0.98 | 0.7 | 0.8 |

Basal-like | 0.76 | 0.98 | N/A | 0.67 | 0.94 |

HER2+/ER- | 0.42 | 0.7 | 0.67 | N/A | 0.46 |

normal-like | 0.83 | 0.8 | 0.94 | 0.46 | N/A |

### Discussion

We proposed a new approach to uncover the transcriptional regulatory networks from microarray gene expression profiles. We discuss next a few distinct features of it.

First, to reflect the fact that a TF only regulates a small number of genes among the whole genome, the loading matrix of the factor model is constrained by a sparse prior [16], which directly reflects our existing knowledge of the particular TF-gene regulation, i.e., if the regulation exists according to prior knowledge, the probability of the corresponding component of the loading matrix to be non-zero is large; or otherwise, very small. The introduction of sparsity significantly constrains the factor model and helps to enable the inference of a set of correlated samples.

To model the samples correlation due to, for instance, cancer subtypes, the samples are modeled by a Dirichlet process mixture (DPM), which imposes clustering effect among samples and can automatically determine the optimal number of clusters from data rather than be predefined in the algorithm. Forth, other types of data, such as ChIP-chip data [39–41] and DNA methylation data [42] can be conveniently integrated with gene expression data [43] under the proposed framework by setting a slightly different prior probabilities to the loading matrix. Integrating additional data types can potentially improve the accuracy of the reconstructed networks. [12].

However, the proposed model is not without shortcomings. First, this model can not yet capture regulations from TFs that are not specified in the prior knowledge database. In reality, it is possible that some TFs that are not specified in the prior knowledge actually regulate the gene transcription. Second, the algorithm may not converge in a reasonable number of iterations on a large data set, thus cannot yet be applied to genome wide data set. Because the model parameters are high dimensional and highly correlated, the speed of convergence may significantly slow down on a large data set [44, 45]. However, such restriction on the size of genes and TFs forces us to focus the analysis on most relevant genes and TFs, making the analysis more targeted and easy to interpret. We demonstrate in section Result, how such analysis can be carried out starting from a whole genome microarray data. With the advancement in ChIP-seq technology and increasing knowledge of TFs biological functions, the proposed model could be applied for a genome-wide study in the future.

*p*, recall

*r*of prior information and the sparsity of the loading matrix

*s*is given, the prior probability of the

*g*-th gene to be a target of the

*l*-th TF

*π*

_{ g }

_{,}

_{ l }can be calculated as follows:

However, the precision or recall of the prior knowledge database are only arbitrarily specified (both 90%). In practice, the quality of prior knowledge should be evaluated first before more reasonable prior probabilities of regulations can be assigned.

## Conclusions

A Bayesian factor model that has sparse loading matrix, non-negative factors, and correlated samples, was proposed to unveil the latent activities of transcription factors and their targeted genes from observed gene mRNA expression profiles. By naturally incorporating the prior knowledge of TF regulated genes, the sparsity constraint of the loading matrix, the non-negativity constraints of TF activities, the regulation coefficients and TF activities can be estimated. A Gibbs sampling solution was proposed and model inference. The effectiveness and validity of the model and the proposed Gibbs sampler were evaluated on simulated systems. The proposed method was applied to the breast cancer microarray data and a TF regulated network for breast cancer data was reconstructed. The inferred TF activities indicates 3 patients clusters, two of which possess significant differences in survival time after treatment. These results demonstrated that the BNFM provides a viable approach to reconstruct TF mediated regulatory networks and estimate TF activities from mRNA expression profiles. The BNFM will be an important tool for not only understanding the transcriptional regulation but also predicting the clinical outcomes of treatment.

## Declarations

### Acknowledgements

This article has been published as part of *Proteome Science* Volume 9 Supplement 1, 2011: Proceedings of the International Workshop on Computational Proteomics. The full contents of the supplement are available online at http://www.proteomesci.com/supplements/9/S1.

## Authors’ Affiliations

## References

- Hobert O:
**Gene regulation by transcription factors and microRNAs.***Science*2008,**319**(5871):1785. 10.1126/science.1151651PubMedView Article - Huang Y, Tienda-Luna I, Wang Y:
**Reverse engineering gene regulatory networks.***Signal Processing Magazine, IEEE*2009,**26:**76–97.View Article - Greenbaum D, Colangelo C, Williams K, Gerstein M:
**Comparing protein abundance and mRNA expression levels on a genomic scale.***Genome Biol*2003,**4**(9):117. 10.1186/gb-2003-4-9-117PubMed CentralPubMedView Article - Gygi S, Rochon Y, Franza B, Aebersold R:
**Correlation between protein and mRNA abundance in yeast.***Molecular and Cellular Biology*1999,**19**(3):1720.PubMed CentralPubMed - Sabatti C, James G:
**Bayesian sparse hidden components analysis for transcription regulation networks.***Bioinformatics*2006,**22**(6):739. 10.1093/bioinformatics/btk017PubMedView Article - Sanguinetti G, Lawrence N, Rattray M:
**Probabilistic inference of transcription factor concentrations and gene-specific regulatory activities.***Bioinformatics*2006,**22**(22):2775. 10.1093/bioinformatics/btl473PubMedView Article - Yu T, Li K:
**Inference of transcriptional regulatory network by two-stage constrained space factor analysis.***Bioinformatics*2005,**21**(21):4033. 10.1093/bioinformatics/bti656PubMedView Article - Boulesteix A, Strimmer K:
**Predicting transcription factor activities from combined analysis of microarray and ChIP data: a partial least squares approach.***Theoretical Biology and Medical Modelling*2005,**2:**23. 10.1186/1742-4682-2-23PubMed CentralPubMedView Article - Kao K, Yang Y, Boscolo R, Sabatti C, Roychowdhury V, Liao J:
**Transcriptome-based determination of multiple transcription regulator activities in Escherichia coli by using network component analysis.***Proceedings of the National Academy of Sciences*2004,**101**(2):641. 10.1073/pnas.0305287101View Article - Meng J, Zhang JM, Qi YA, Chen Y, Huang Y:
**Uncovering Transcriptional Regulatory Networks by Sparse Bayesian Factor Model.***Eurasip Journal On Advances In Signal Processing*2010. - Matys V, Fricke E, Geffers R, Gossling E, Haubrock M, Hehl R, Hornischer K, Karas D, Kel A, Kel-Margoulis O,
*et al*.:**TRANSFAC (R): transcriptional regulation, from patterns to profiles.***Nucleic acids research*2003,**31:**374. 10.1093/nar/gkg108PubMed CentralPubMedView Article - Ideker Trey, JHL Dutkowski:
**Boosting Signal-to-Noise in Complex Biology: Prior Knowledge Is Power.***Cell*2011,**144**(6):860–863. 10.1016/j.cell.2011.03.007PubMed CentralPubMedView Article - Qi Q, Zhao Y, Li M, Simon R:
**Non-negative matrix factorization of gene expression profiles: a plug-in for BRB-ArrayTools.***Bioinformatics*2009,**25**(4):545. 10.1093/bioinformatics/btp009PubMedView Article - Hoyer P:
**Non-negative matrix factorization with sparseness constraints.***The Journal of Machine Learning Research*2004,**5:**1469. - Brunet J, Tamayo P, Golub T, Mesirov J:
**Metagenes and molecular pattern discovery using matrix factorization.***Proceedings of the National Academy of Sciences of the United States of America*2004,**101**(12):4164. 10.1073/pnas.0308531101PubMed CentralPubMedView Article - Carvalho C, Chang J, Lucas J, Nevins J, Wang Q, West M:
**High-dimensional sparse factor modeling: Applications in gene expression genomics.***Journal of the American Statistical Association*2008,**103**(484):1438–1456. 10.1198/016214508000000869PubMed CentralPubMedView Article - Sudderth E:
**Graphical models for visual object recognition and tracking.***PhD thesis*. Massachusetts Institute of Technology; 2006. - Ferguson T:
**A Bayesian analysis of some nonparametric problems.***The annals of statistics*1973,**1**(2):209–230. 10.1214/aos/1176342360View Article - Socci N, Lee D, Sebastian Seung H:
**The rectified Gaussian distribution.***Advances in Neural Information Processing Systems*1998, 350–356. - Cui X, Churchill G:
**Statistical tests for differential expression in cDNA microarray experiments.***Genome Biol*2003,**4**(4):210. 10.1186/gb-2003-4-4-210PubMed CentralPubMedView Article - Wong C:
**Differential Expression and Annotation.**2009. - Wilson D, Charoensawan V, Kummerfeld S, Teichmann S:
**DBD-taxonomically broad transcription factor predictions: new content and functionality.***Nucleic Acids Research*2008,**36**(Database issue):D88.PubMed CentralPubMed - Tipping M, Bishop C:
**Probabilistic principal component analysis.***Journal of the Royal Statistical Society: Series B (Statistical Methodology)*1999,**61**(3):611–622. 10.1111/1467-9868.00196View Article - Gelman A, Carlin J, Stern H, Rubin D:
**Bayesian data analysis.***London, Glasgow, et al*1995. - Thompson W, Newberg L, Conlan S, McCue L, Lawrence C:
**The Gibbs centroid sampler.***Nucleic Acids Research*2007. - Van Rijsbergen C:
**Foundation of evaluation.***Journal of Documentation*1974,**30**(4):365–373. 10.1108/eb026584View Article - Bagga A, Baldwin B:
**Entity-based cross-document coreferencing using the vector space model.**In*Proceedings of the 17th international conference on Computational linguistics-Volume 1*. Association for Computational Linguistics Morristown, NJ, USA; 1998:79–85.View Article - Amigó E, Gonzalo J, Artiles J, Verdejo F:
**A comparison of extrinsic clustering evaluation metrics based on formal constraints.***Information Retrieval*2009,**12**(4):461–486. 10.1007/s10791-008-9066-8View Article - Hoadley K, Weigman V, Fan C, Sawyer L, He X, Troester M, Sartor C, Rieger-House T, Bernard P, Carey L,
*et al*.:**EGFR associated expression profiles vary with breast tumor subtype.***BMC genomics*2007,**8:**258. 10.1186/1471-2164-8-258PubMed CentralPubMedView Article - Mullins M, Perreard L, Quackenbush J, Gauthier N, Bayer S, Ellis M, Parker J, Perou C, Szabo A, Bernard P:
**Agreement in breast cancer classification between microarray and quantitative reverse transcription PCR from fresh-frozen and formalin-fixed, paraffin-embedded tissues.***Clinical chemistry*2007,**53**(7):1273. 10.1373/clinchem.2006.083725PubMedView Article - Herschkowitz J, Simin K, Weigman V, Mikaelian I, Usary J, Hu Z, Rasmussen K, Jones L, Assefnia S, Chandrasekharan S,
*et al*.:**Identification of conserved gene expression features between murine mammary carcinoma models and human breast tumors.***Genome biology*2007,**8**(5):R76. 10.1186/gb-2007-8-5-r76PubMed CentralPubMedView Article - Herschkowitz J, He X, Fan C, Perou C:
**The functional loss of the retinoblastoma tumour suppressor is a common event in basal-like and luminal B breast carcinomas.***Breast Cancer Res*2008,**10**(5):R75. 10.1186/bcr2142PubMed CentralPubMedView Article - Perou C, Sørlie T, Eisen M, van de Rijn M, Jeffrey S, Rees C, Pollack J, Ross D, Johnsen H, Akslen L,
*et al*.:**Molecular portraits of human breast tumours.***Nature*2000,**406**(6797):747–752. 10.1038/35021093PubMedView Article - Sørlie T, Perou C, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen M, Van De Rijn M, Jeffrey S,
*et al*.:**Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications.***Proceedings of the National Academy of Sciences of the United States of America*2001,**98**(19):10869. 10.1073/pnas.191367098PubMed CentralPubMedView Article - Sørlie T, Tibshirani R, Parker J, Hastie T, Marron J, Nobel A, Deng S, Johnsen H, Pesich R, Geisler S,
*et al*.:**Repeated observation of breast tumor subtypes in independent gene expression data sets.***Proceedings of the National Academy of Sciences of the United States of America*2003,**100**(14):8418. 10.1073/pnas.0932692100PubMed CentralPubMedView Article - Shai R, Shi T, Kremen T, Horvath S, Liau L, Cloughesy T, Mischel P, Nelson S:
**Gene expression profiling identifies molecular subtypes of gliomas.***Oncogene*2003,**22**(31):4918–4923. 10.1038/sj.onc.1206753PubMedView Article - Kim P, Tidor B:
**Subsystem identification through dimensionality reduction of large-scale gene expression data.***Genome Research*2003,**13**(7):1706. 10.1101/gr.903503PubMed CentralPubMedView Article - Li T, Ding C:
**The relationships among various nonnegative matrix factorization methods for clustering.***Data Mining, 2006.ICDM’06. Sixth International Conference on*2006, 362–371. - Lieb J, Liu X, Botstein D, Brown P:
**Promoter-specific binding of Rap1 revealed by genome-wide maps of protein-DNA association.***Nature genetics*2001,**28**(4):327–334. 10.1038/ng569PubMedView Article - Iyer V, Horak C, Scafe C, Botstein D, Snyder M, Brown P:
**Genomic binding sites of the yeast cell-cycle transcription factors SBF and MBF.***Nature*2001,**409**(6819):533–538. 10.1038/35054095PubMedView Article - Ren B, Robert F, Wyrick J, Aparicio O, Jennings E, Simon I, Zeitlinger J, Schreiber J, Hannett N, Kanin E,
*et al*.:**Genome-wide location and function of DNA binding proteins.***Science’s STKE*2000,**290**(5500):2306. - Jaenisch R, Bird A:
**Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals.***Nature genetics*2003,**33:**245–254. 10.1038/ng1089PubMedView Article - Tasheva E, Klocke B, Conrad G:
**Analysis of transcriptional regulation of the small leucine rich proteoglycans.***Mol Vis*2004,**10:**758–772.PubMed - Justel A:
**Gibbs sampling will fail in outlier problems with strong masking.***Journal of Computational and Graphical Statistics*1996,**5**(2):176–189. 10.2307/1390779 - Borgs C, Chayes J, Frieze A, Kim J, Tetali P, Vigoda E, Vu V:
**Torpid mixing of some Monte Carlo Markov chain algorithms in statistical physics.***ANNUAL SYMPOSIUM ON FOUNDATIONS OF COMPUTER SCIENCE, Volume 40*1999, 218–229.

## Copyright

This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.