Enzymatic Polymerization of High Molecular Weight DNA

The use of DNA as a polymeric building material transcends its function in biology and is exciting in bionanotechnology for applications ranging from biosensing, to diagnostics, and to targeted drug delivery. These applications are enabled by DNA’s unique structural and chemical properties, embodied as a directional polyanion that exhibits molecular recognition capabilities. Hence, the efficient and precise synthesis of high molecular weight DNA materials has become key to advance DNA bionanotechnology. Current synthesis methods largely rely on either solid phase chemical synthesis or template-dependent polymerase amplification. The inherent step-by-step fashion of solid phase synthesis limits the length of the resulting DNA to typically less than 150 nucleotides. In contrast, polymerase based enzymatic synthesis methods (e.g., polymerase chain reaction) are not limited by product length, but require a DNA template to guide the synthesis. Furthermore, advanced DNA bionanotechnology requires tailorable structural and self-assembly properties. Current synthesis methods, however, often involve multiple conjugating reactions and extensive purification steps.

The research described in this dissertation aims to develop a facile method to synthesize high molecular weight, single stranded DNA (or polynucleotide) with versatile functionalities. We exploit the ability of a template-independent DNA polymerase−terminal deoxynucleotidyl transferase (TdT) to catalyze the polymerization of 2’-deoxyribonucleoside 5’-triphosphates (dNTP, monomer) from the 3’-hydroxyl group of an oligodeoxyribonucleotide (initiator). We termed this enzymatic synthesis method: TdT catalyzed enzymatic polymerization, or TcEP.

Specifically, this dissertation is structured to address three specific research aims. With the objective to generate high molecular weight polynucleotides, Specific Aim 1 studies the reaction kinetics of TcEP by investigating the polymerization of 2’-deoxythymidine 5’-triphosphates (monomer) from the 3’-hydroxyl group of oligodeoxyribothymidine (initiator) using in situ 1H NMR and fluorescent gel electrophoresis. We found that TcEP kinetics follows the “living” chain-growth polycondensation mechanism, and like in “living” polymerizations, the molecular weight of the final product is determined by the starting molar ratio of monomer to initiator. The distribution of the molecular weight is crucially influenced by the molar ratio of initiator to TdT. We developed a reaction kinetics model that allows us to quantitatively describe the reaction and predict the molecular weight of the reaction products.

Specific Aim 2 further explores TcEP’s ability to transcend homo-polynucleotide synthesis by varying the choices of initiators and monomers. We investigated the effects of initiator length and sequence on TcEP, and found that the minimum length of an effective initiator should be 10 nucleotides and that the formation of secondary structures close to the 3’-hydroxyl group can impede the polymerization reaction. We also demonstrated TcEP’s capacity to incorporate a wide range of unnatural dNTPs into the growing chain, such as, hydrophobic fluorescent dNTP and fluoro modified dNTP. By harnessing the encoded nucleotide sequence of an initiator and the chemical diversity of monomers, TcEP enables us to introduce molecular recognition capabilities and chemical functionalities on the 5’-terminus and 3’-terminus, respectively.

Building on TcEP’s synthesis capacities, in Specific Aim 3 we invented a two-step strategy to synthesize diblock amphiphilic polynucleotides, in which the first, hydrophilic block serves as a macro-initiator for the growth of the second block, comprised of natural and/or unnatural nucleotides. By tuning the hydrophilic length, we synthesized the amphiphilic diblock polynucleotides that can self-assemble into micellar structures ranging from star-like to crew-cut morphologies. The observed self-assembly behaviors agree with predictions from dissipative particle dynamics simulations as well as scaling law for polyelectrolyte block copolymers.

In summary, we developed an enzymatic synthesis method (i.e., TcEP) that enables the facile synthesis of high molecular weight polynucleotides with low polydispersity. Although we can control the nucleotide sequence only to a limited extent, TcEP offers a method to integrate an oligodeoxyribonucleotide with specific sequence at the 5’-terminus and to incorporate functional groups along the growing chains simultaneously. Additionally, we used TcEP to synthesize amphiphilic polynucleotides that display self-assemble ability. We anticipate that our facile synthesis method will not only advance molecular biology, but also invigorate materials science and bionanotechnology.

Distinct functions of POT1 at telomeres.

The mammalian protein POT1 binds to telomeric single-stranded DNA (ssDNA), protecting chromosome ends from being detected as sites of DNA damage. POT1 is composed of an N-terminal ssDNA-binding domain and a C-terminal protein interaction domain. With regard to the latter, POT1 heterodimerizes with the protein TPP1 to foster binding to telomeric ssDNA in vitro and binds the telomeric double-stranded-DNA-binding protein TRF2. We sought to determine which of these functions-ssDNA, TPP1, or TRF2 binding-was required to protect chromosome ends from being detected as DNA damage. Using separation-of-function POT1 mutants deficient in one of these three activities, we found that binding to TRF2 is dispensable for protecting telomeres but fosters robust loading of POT1 onto telomeric chromatin. Furthermore, we found that the telomeric ssDNA-binding activity and binding to TPP1 are required in cis for POT1 to protect telomeres. Mechanistically, binding of POT1 to telomeric ssDNA and association with TPP1 inhibit the localization of RPA, which can function as a DNA damage sensor, to telomeres.

Microbial inactivation of Pseudomonas putida and Pichia pastoris using gene silencing.

Antisense deoxyoligonucleotide (ASO) gene silencing was investigated as a potential disinfection tool for industrial and drinking water treatment application. ASOs bind with their reverse complementary mRNA transcripts thereby blocking protein translation. While ASO silencing has mainly been studied in medicine, it may be useful for modulating gene expression and inactivating microorganisms in environmental applications. In this proof of concept work, gene targets were sh ble (zeocin resistance) and todE (catechol-2,3-dioxygenase) in Pichia pastoris and npt (kanamycin resistance) in Pseudomonas putida. A maximum 0.5-fold decrease in P. pastoris cell numbers was obtained following a 120 min incubation with single-stranded DNA (ssDNA) concentrations ranging from 0.2 to 200 nM as compared to the no ssDNA control. In P. putida, a maximum 5.2-fold decrease was obtained after 90 min with 400 nM ssDNA. While the silencing efficiencies varied for the 25 targets tested, these results suggest that protein activity as well as microbial growth can be altered using ASO gene silencing-based tools. If successful, this technology has the potential to eliminate some of the environmental and health issues associated with the use of strong chemical biocides. However, prior to its dissemination, more research is needed to increase silencing efficiency and develop effective delivery methods.

Suppression of DNA-damage checkpoint signaling by Rsk-mediated phosphorylation of Mre11.

Ataxia telangiectasia mutant (ATM) is an S/T-Q-directed kinase that is critical for the cellular response to double-stranded breaks (DSBs) in DNA. Following DNA damage, ATM is activated and recruited by the MRN protein complex [meiotic recombination 11 (Mre11)/DNA repair protein Rad50/Nijmegen breakage syndrome 1 proteins] to sites of DNA damage where ATM phosphorylates multiple substrates to trigger cell-cycle arrest. In cancer cells, this regulation may be faulty, and cell division may proceed even in the presence of damaged DNA. We show here that the ribosomal s6 kinase (Rsk), often elevated in cancers, can suppress DSB-induced ATM activation in both Xenopus egg extracts and human tumor cell lines. In analyzing each step in ATM activation, we have found that Rsk targets loading of MRN complex components onto DNA at DSB sites. Rsk can phosphorylate the Mre11 protein directly at S676 both in vitro and in intact cells and thereby can inhibit the binding of Mre11 to DNA with DSBs. Accordingly, mutation of S676 to Ala can reverse inhibition of the response to DSBs by Rsk. Collectively, these data point to Mre11 as an important locus of Rsk-mediated checkpoint inhibition acting upstream of ATM activation.

Regulation of DNA Double Strand Break Response

To ensure genomic integrity, dividing cells implement multiple checkpoint pathways during the course of the cell cycle. In response to DNA damage, cells may either halt the progression of the cycle (cell cycle arrest) or undergo apoptosis. This choice depends on the extent of damage and the cell's capacity for DNA repair. Cell cycle arrest induced by double-stranded DNA breaks relies on the activation of the ataxia-telangiectasia (ATM) protein kinase, which phosphorylates cell cycle effectors (e.g., Chk2 and p53) to inhibit cell cycle progression. ATM is an S/T-Q directed kinase that is critical for the cellular response to double-stranded DNA breaks. Following DNA damage, ATM is activated and recruited to sites of DNA damage by the MRN protein complex (Mre11-Rad50-Nbs1 proteins) where ATM phosphorylates multiple substrates to trigger a cell cycle arrest. In cancer cells, this regulation may be faulty and cell division may proceed even in the presence of damaged DNA. We show here that the RSK kinase, often elevated in cancers, can suppress DSB-induced ATM activation in both Xenopus egg extracts and human tumor cell lines. In analyzing each step in ATM activation, we have found that RSK disrupts the binding of the MRN complex to DSB DNA. RSK can directly phosphorylate the Mre11 protein at Ser 676 both in vitro and in intact cells and can thereby inhibit loading of Mre11 onto DSB DNA. Accordingly, mutation of Ser 676 to Ala can reverse inhibition of the DSB response by RSK. Collectively, these data point to Mre11 as an important locus of RSK-mediated checkpoint inhibition acting upstream of ATM activation.

The phosphorylation of Mre11 on Ser 676 is antagonized by phosphatases. Here, we screened for phosphatases that target this site and identified PP5 as a candidate. This finding is consistent with the fact that PP5 is required for the ATM-mediated DNA damage response, indicating that PP5 may promote DSB-induced, ATM-dependent DNA damage response by targeting Mre11 upstream of ATM.

Programming Molecular Devices using Nucleic Acid Hairpins

Nucleic Acid hairpins have been a subject of study for the last four decades. They are composed of single strand that is

hybridized to itself, and the central section forming an unhybridized loop. In nature, they stabilize single stranded RNA, serve as nucleation

sites for RNA folding, protein recognition signals, mRNA localization and regulation of mRNA degradation. On the other hand,

DNA hairpins in biological contexts have been studied with respect to forming cruciform structures that can regulate gene expression.

The use of DNA hairpins as fuel for synthetic molecular devices, including locomotion, was proposed and experimental demonstrated in 2003. They

were interesting because they bring to the table an on-demand energy/information supply mechanism.

The energy/information is hidden (from hybridization) in the hairpin’s loop, until required.

The energy/information is harnessed by opening the stem region, and exposing the single stranded loop section.

The loop region is now free for possible hybridization and help move the system into a thermodynamically favourable state.

The hidden energy and information coupled with

programmability provides another functionality, of selectively choosing what reactions to hide and

what reactions to allow to proceed, that helps develop a topological sequence of events.

Hairpins have been utilized as a source of fuel for many different DNA devices. In this thesis, we program four different

molecular devices using DNA hairpins, and experimentally validate them in the

laboratory. 1) The first device: A

novel enzyme-free autocatalytic self-replicating system composed entirely of DNA that operates isothermally. 2) The second

device: Time-Responsive Circuits using DNA have two properties: a) asynchronous: the final output is always correct

regardless of differences in the arrival time of different inputs.

b) renewable circuits which can be used multiple times without major degradation of the gate motifs

(so if the inputs change over time, the DNA-based circuit can re-compute the output correctly based on the new inputs).

3) The third device: Activatable tiles are a theoretical extension to the Tile assembly model that enhances

its robustness by protecting the sticky sides of tiles until a tile is partially incorporated into a growing assembly.

4) The fourth device: Controlled Amplification of DNA catalytic system: a device such that the amplification

of the system does not run uncontrollably until the system runs out of fuel, but instead achieves a finite

amount of gain.

Nucleic acid circuits with the ability

to perform complex logic operations have many potential practical applications, for example the ability to achieve point of care diagnostics.

We discuss the designs of our DNA Hairpin molecular devices, the results we have obtained, and the challenges we have overcome

to make these truly functional.

The upstream enhancer elements of the G6PC promoter are critical for optimal G6PC expression in murine glycogen storage disease type Ia.

Glycogen storage disease type-Ia (GSD-Ia) patients deficient in glucose-6-phosphatase-α (G6Pase-α or G6PC) manifest impaired glucose homeostasis characterized by fasting hypoglycemia, growth retardation, hepatomegaly, nephromegaly, hyperlipidemia, hyperuricemia, and lactic acidemia. Two efficacious recombinant adeno-associated virus pseudotype 2/8 (rAAV8) vectors expressing human G6Pase-α have been independently developed. One is a single-stranded vector containing a 2864-bp of the G6PC promoter/enhancer (rAAV8-GPE) and the other is a double-stranded vector containing a shorter 382-bp minimal G6PC promoter/enhancer (rAAV8-miGPE). To identify the best construct, a direct comparison of the rAAV8-GPE and the rAAV8-miGPE vectors was initiated to determine the best vector to take forward into clinical trials. We show that the rAAV8-GPE vector directed significantly higher levels of hepatic G6Pase-α expression, achieved greater reduction in hepatic glycogen accumulation, and led to a better toleration of fasting in GSD-Ia mice than the rAAV8-miGPE vector. Our results indicated that additional control elements in the rAAV8-GPE vector outweigh the gains from the double-stranded rAAV8-miGPE transduction efficiency, and that the rAAV8-GPE vector is the current choice for clinical translation in human GSD-Ia.

Development and Characterization of Monovalent and Bivalent RNA Aptamers Targeting the Common Pathway of Coagulation

Anticoagulant agents are commonly used drugs to reduce blood coagulation in acute and chronic clinical settings. Many of these drugs target the common pathway of coagulation because it is critical for thrombin generation and disruption of this portion of the pathway has profound effects on the hemostatic process. Currently available drugs for these indications struggle with balancing desired activity with immunogenicity and poor reversibility or irreversibility in the event of hemorrhage. While improvements are being made with the current drugs, new drugs with better therapeutic indices are needed for surgical intervention and chronic indications to prevent thrombosis from occurring.

A class of therapeutics known as aptamers may be able to meet the need for safer anticoagulant agents. Aptamer are short single-stranded RNA oligonucleotides that adopt specific secondary and tertiary structures based upon their sequence. They can be generated to both enzymes and cofactors because they derive their inhibitory activity by blocking protein-protein interactions, rather than active site inhibition. They inhibit their target proteins with a high level of specificity and bind with high affinity to their target. Additionally, they can be reversed using two different antidote approaches, specific oligonucleotide antidotes, or with cationic, “universal” antidotes. The reversal of their activity is both rapid and durable.

The ability of aptamers to be generated to cofactors has been conclusively proven by generating an aptamer targeting the common pathway coagulation cofactor, Factor V (FV). We developed two aptamers with anticoagulant ability that bind to both FV and FVa, the active cofactor. Both aptamers were truncated to smaller functional sizes and had specific point mutant aptamers developed for use as controls. The anticoagulant activity of both aptamer-mutant pairs was characterized using plasma-based clotting assays and whole blood assays. The mechanism of action resulting in anticoagulant activity was assessed for one aptamer. The aptamer was found to block FVa docking to membrane surfaces, a mechanism not previously observed in any of our other anticoagulant aptamers.

To explore development of aptamers as anticoagulant agents targeting the common pathway for surgical interventions, we fused two anticoagulant aptamers targeting Factor X and prothrombin into a single molecule. The bivalent aptamer was truncated to a minimal size while maintaining robust anticoagulant activity. Characterization of the bivalent aptamer in plasma-based clotting assays indicated we had generated a very robust anticoagulant therapeutic. Furthermore, we were able to simultaneously reverse the activity of both aptamers with a single oligonucleotide antidote. This rapid and complete reversal of anticoagulant activity is not available in the antithrombotic agents currently used in surgery.

Separating DNA with different topologies by atomic force microscopy in comparison with gel electrophoresis.

Atomic force microscopy, which is normally used for DNA imaging to gain qualitative results, can also be used for quantitative DNA research, at a single-molecular level. Here, we evaluate the performance of AFM imaging specifically for quantifying supercoiled and relaxed plasmid DNA fractions within a mixture, and compare the results with the bulk material analysis method, gel electrophoresis. The advantages and shortcomings of both methods are discussed in detail. Gel electrophoresis is a quick and well-established quantification method. However, it requires a large amount of DNA, and needs to be carefully calibrated for even slightly different experimental conditions for accurate quantification. AFM imaging is accurate, in that single DNA molecules in different conformations can be seen and counted. When used carefully with necessary correction, both methods provide consistent results. Thus, AFM imaging can be used for DNA quantification, as an alternative to gel electrophoresis.

Mitochondrial DNA polymorphism A4917G is independently associated with age-related macular degeneration.

The objective of this study was to determine if MTND2*LHON4917G (4917G), a specific non-synonymous polymorphism in the mitochondrial genome previously associated with neurodegenerative phenotypes, is associated with increased risk for age-related macular degeneration (AMD). A preliminary study of 393 individuals (293 cases and 100 controls) ascertained at Vanderbilt revealed an increased occurrence of 4917G in cases compared to controls (15.4% vs.9.0%, p = 0.11). Since there was a significant age difference between cases and controls in this initial analysis, we extended the study by selecting Caucasian pairs matched at the exact age at examination. From the 1547 individuals in the Vanderbilt/Duke AMD population association study (including 157 in the preliminary study), we were able to match 560 (280 cases and 280 unaffected) on exact age at examination. This study population was genotyped for 4917G plus specific AMD-associated nuclear genome polymorphisms in CFH, LOC387715 and ApoE. Following adjustment for the listed nuclear genome polymorphisms, 4917G independently predicts the presence of AMD (OR = 2.16, 95%CI 1.20-3.91, p = 0.01). In conclusion, a specific mitochondrial polymorphism previously implicated in other neurodegenerative phenotypes (4917G) appears to convey risk for AMD independent of recently discovered nuclear DNA polymorphisms.

Transmission of single HIV-1 genomes and dynamics of early immune escape revealed by ultra-deep sequencing.

We used ultra-deep sequencing to obtain tens of thousands of HIV-1 sequences from regions targeted by CD8+ T lymphocytes from longitudinal samples from three acutely infected subjects, and modeled viral evolution during the critical first weeks of infection. Previous studies suggested that a single virus established productive infection, but these conclusions were tempered because of limited sampling; now, we have greatly increased our confidence in this observation through modeling the observed earliest sample diversity based on vastly more extensive sampling. Conventional sequencing of HIV-1 from acute/early infection has shown different patterns of escape at different epitopes; we investigated the earliest escapes in exquisite detail. Over 3-6 weeks, ultradeep sequencing revealed that the virus explored an extraordinary array of potential escape routes in the process of evading the earliest CD8 T-lymphocyte responses--using 454 sequencing, we identified over 50 variant forms of each targeted epitope during early immune escape, while only 2-7 variants were detected in the same samples via conventional sequencing. In contrast to the diversity seen within epitopes, non-epitope regions, including the Envelope V3 region, which was sequenced as a control in each subject, displayed very low levels of variation. In early infection, in the regions sequenced, the consensus forms did not have a fitness advantage large enough to trigger reversion to consensus amino acids in the absence of immune pressure. In one subject, a genetic bottleneck was observed, with extensive diversity at the second time point narrowing to two dominant escape forms by the third time point, all within two months of infection. Traces of immune escape were observed in the earliest samples, suggesting that immune pressure is present and effective earlier than previously reported; quantifying the loss rate of the founder virus suggests a direct role for CD8 T-lymphocyte responses in viral containment after peak viremia. Dramatic shifts in the frequencies of epitope variants during the first weeks of infection revealed a complex interplay between viral fitness and immune escape.

Association between DNA damage response and repair genes and risk of invasive serous ovarian cancer.

BACKGROUND: We analyzed the association between 53 genes related to DNA repair and p53-mediated damage response and serous ovarian cancer risk using case-control data from the North Carolina Ovarian Cancer Study (NCOCS), a population-based, case-control study. METHODS/PRINCIPAL FINDINGS: The analysis was restricted to 364 invasive serous ovarian cancer cases and 761 controls of white, non-Hispanic race. Statistical analysis was two staged: a screen using marginal Bayes factors (BFs) for 484 SNPs and a modeling stage in which we calculated multivariate adjusted posterior probabilities of association for 77 SNPs that passed the screen. These probabilities were conditional on subject age at diagnosis/interview, batch, a DNA quality metric and genotypes of other SNPs and allowed for uncertainty in the genetic parameterizations of the SNPs and number of associated SNPs. Six SNPs had Bayes factors greater than 10 in favor of an association with invasive serous ovarian cancer. These included rs5762746 (median OR(odds ratio)(per allele) = 0.66; 95% credible interval (CI) = 0.44-1.00) and rs6005835 (median OR(per allele) = 0.69; 95% CI = 0.53-0.91) in CHEK2, rs2078486 (median OR(per allele) = 1.65; 95% CI = 1.21-2.25) and rs12951053 (median OR(per allele) = 1.65; 95% CI = 1.20-2.26) in TP53, rs411697 (median OR (rare homozygote) = 0.53; 95% CI = 0.35 - 0.79) in BACH1 and rs10131 (median OR( rare homozygote) = not estimable) in LIG4. The six most highly associated SNPs are either predicted to be functionally significant or are in LD with such a variant. The variants in TP53 were confirmed to be associated in a large follow-up study. CONCLUSIONS/SIGNIFICANCE: Based on our findings, further follow-up of the DNA repair and response pathways in a larger dataset is warranted to confirm these results.

Comparative analyses of seven algorithms for copy number variant identification from single nucleotide polymorphism arrays.

Determination of copy number variants (CNVs) inferred in genome wide single nucleotide polymorphism arrays has shown increasing utility in genetic variant disease associations. Several CNV detection methods are available, but differences in CNV call thresholds and characteristics exist. We evaluated the relative performance of seven methods: circular binary segmentation, CNVFinder, cnvPartition, gain and loss of DNA, Nexus algorithms, PennCNV and QuantiSNP. Tested data included real and simulated Illumina HumHap 550 data from the Singapore cohort study of the risk factors for Myopia (SCORM) and simulated data from Affymetrix 6.0 and platform-independent distributions. The normalized singleton ratio (NSR) is proposed as a metric for parameter optimization before enacting full analysis. We used 10 SCORM samples for optimizing parameter settings for each method and then evaluated method performance at optimal parameters using 100 SCORM samples. The statistical power, false positive rates, and receiver operating characteristic (ROC) curve residuals were evaluated by simulation studies. Optimal parameters, as determined by NSR and ROC curve residuals, were consistent across datasets. QuantiSNP outperformed other methods based on ROC curve residuals over most datasets. Nexus Rank and SNPRank have low specificity and high power. Nexus Rank calls oversized CNVs. PennCNV detects one of the fewest numbers of CNVs.

Computational Inference of Genome-Wide Protein-DNA Interactions Using High-Throughput Genomic Data

Transcriptional regulation has been studied intensively in recent decades. One important aspect of this regulation is the interaction between regulatory proteins, such as transcription factors (TF) and nucleosomes, and the genome. Different high-throughput techniques have been invented to map these interactions genome-wide, including ChIP-based methods (ChIP-chip, ChIP-seq, etc.), nuclease digestion methods (DNase-seq, MNase-seq, etc.), and others. However, a single experimental technique often only provides partial and noisy information about the whole picture of protein-DNA interactions. Therefore, the overarching goal of this dissertation is to provide computational developments for jointly modeling different experimental datasets to achieve a holistic inference on the protein-DNA interaction landscape.

We first present a computational framework that can incorporate the protein binding information in MNase-seq data into a thermodynamic model of protein-DNA interaction. We use a correlation-based objective function to model the MNase-seq data and a Markov chain Monte Carlo method to maximize the function. Our results show that the inferred protein-DNA interaction landscape is concordant with the MNase-seq data and provides a mechanistic explanation for the experimentally collected MNase-seq fragments. Our framework is flexible and can easily incorporate other data sources. To demonstrate this flexibility, we use prior distributions to integrate experimentally measured protein concentrations.

We also study the ability of DNase-seq data to position nucleosomes. Traditionally, DNase-seq has only been widely used to identify DNase hypersensitive sites, which tend to be open chromatin regulatory regions devoid of nucleosomes. We reveal for the first time that DNase-seq datasets also contain substantial information about nucleosome translational positioning, and that existing DNase-seq data can be used to infer nucleosome positions with high accuracy. We develop a Bayes-factor-based nucleosome scoring method to position nucleosomes using DNase-seq data. Our approach utilizes several effective strategies to extract nucleosome positioning signals from the noisy DNase-seq data, including jointly modeling data points across the nucleosome body and explicitly modeling the quadratic and oscillatory DNase I digestion pattern on nucleosomes. We show that our DNase-seq-based nucleosome map is highly consistent with previous high-resolution maps. We also show that the oscillatory DNase I digestion pattern is useful in revealing the nucleosome rotational context around TF binding sites.

Finally, we present a state-space model (SSM) for jointly modeling different kinds of genomic data to provide an accurate view of the protein-DNA interaction landscape. We also provide an efficient expectation-maximization algorithm to learn model parameters from data. We first show in simulation studies that the SSM can effectively recover underlying true protein binding configurations. We then apply the SSM to model real genomic data (both DNase-seq and MNase-seq data). Through incrementally increasing the types of genomic data in the SSM, we show that different data types can contribute complementary information for the inference of protein binding landscape and that the most accurate inference comes from modeling all available datasets.

This dissertation provides a foundation for future research by taking a step toward the genome-wide inference of protein-DNA interaction landscape through data integration.