Reverse Engineering Gene Networks Research Articles

Algorithmic reconstruction of glioblastoma network complexity

SummaryGlioblastoma is a complex disease that is difficult to treat. Network and data science offer alternative approaches to classical bioinformatics pipelines to study gene expression patterns from single-cell datasets, helping to distinguish genes associated with the control of differentiation and aggression. To identify the key molecular regulators of the networks driving glioblastoma/GSC and predict their cell fate dynamics, we applied a host of data theoretic techniques to gene expression patterns from pediatric and adult glioblastoma, and adult glioma-derived stem cells (GSCs). We identified eight transcription factors (OLIG1/2, TAZ, GATA2, FOXG1, SOX6, SATB2, and YY1) and four signaling genes (ATL3, MTSS1, EMP1, and TPT1) as coordinators of cell state transitions and, thus, clinically targetable putative factors differentiating pediatric and adult glioblastomas from adult GSCs. Our study provides strong evidence of complex systems approaches for inferring complex dynamics from reverse-engineering gene networks, bolstering the search for new clinically relevant targets in glioblastoma.

Reverse engineering gene networks using global-local shrinkage rules.

Inferring gene regulatory networks from high-throughput 'omics' data has proven to be a computationally demanding task of critical importance. Frequently, the classical methods break down owing to the curse of dimensionality, and popular strategies to overcome this are typically based on regularized versions of the classical methods. However, these approaches rely on loss functions that may not be robust and usually do not allow for the incorporation of prior information in a straightforward way. Fully Bayesian methods are equipped to handle both of these shortcomings quite naturally, and they offer the potential for improvements in network structure learning. We propose a Bayesian hierarchical model to reconstruct gene regulatory networks from time-series gene expression data, such as those common in perturbation experiments of biological systems. The proposed methodology uses global-local shrinkage priors for posterior selection of regulatory edges and relaxes the common normal likelihood assumption in order to allow for heavy-tailed data, which were shown in several of the cited references to severely impact network inference. We provide a sufficient condition for posterior propriety and derive an efficient Markov chain Monte Carlo via Gibbs sampling in the electronic supplementary material. We describe a novel way to detect multiple scales based on the corresponding posterior quantities. Finally, we demonstrate the performance of our approach in a simulation study and compare it with existing methods on real data from a T-cell activation study.

Reverse-engineering of gene networks for regulating early blood development from single-cell measurements

BackgroundRecent advances in omics technologies have raised great opportunities to study large-scale regulatory networks inside the cell. In addition, single-cell experiments have measured the gene and protein activities in a large number of cells under the same experimental conditions. However, a significant challenge in computational biology and bioinformatics is how to derive quantitative information from the single-cell observations and how to develop sophisticated mathematical models to describe the dynamic properties of regulatory networks using the derived quantitative information.MethodsThis work designs an integrated approach to reverse-engineer gene networks for regulating early blood development based on singel-cell experimental observations. The wanderlust algorithm is initially used to develop the pseudo-trajectory for the activities of a number of genes. Since the gene expression data in the developed pseudo-trajectory show large fluctuations, we then use Gaussian process regression methods to smooth the gene express data in order to obtain pseudo-trajectories with much less fluctuations. The proposed integrated framework consists of both bioinformatics algorithms to reconstruct the regulatory network and mathematical models using differential equations to describe the dynamics of gene expression.ResultsThe developed approach is applied to study the network regulating early blood cell development. A graphic model is constructed for a regulatory network with forty genes and a dynamic model using differential equations is developed for a network of nine genes. Numerical results suggests that the proposed model is able to match experimental data very well. We also examine the networks with more regulatory relations and numerical results show that more regulations may exist. We test the possibility of auto-regulation but numerical simulations do not support the positive auto-regulation. In addition, robustness is used as an importantly additional criterion to select candidate networks.ConclusionThe research results in this work shows that the developed approach is an efficient and effective method to reverse-engineer gene networks using single-cell experimental observations.

Accelerated parallel algorithm for gene network reverse engineering

BackgroundThe Algorithm for the Reconstruction of Accurate Cellular Networks (ARACNE) represents one of the most effective tools to reconstruct gene regulatory networks from large-scale molecular profile datasets. However, previous implementations require intensive computing resources and, in some cases, restrict the number of samples that can be used. These issues can be addressed elegantly in a GPU computing framework, where repeated mathematical computation can be done efficiently, but requires extensive redesign to apply parallel computing techniques to the original serial algorithm, involving detailed optimization efforts based on a deep understanding of both hardware and software architecture.ResultHere, we present an accelerated parallel implementation of ARACNE (GPU-ARACNE). By taking advantage of multi-level parallelism and the Compute Unified Device Architecture (CUDA) parallel kernel-call library, GPU-ARACNE successfully parallelizes a serial algorithm and simplifies the user experience from multi-step operations to one step. Using public datasets on comparable hardware configurations, we showed that GPU-ARACNE is faster than previous implementations and is able to reconstruct equally valid gene regulatory networks.ConclusionGiven that previous versions of ARACNE are extremely resource demanding, either in computational time or in hardware investment, GPU-ARACNE is remarkably valuable for researchers who need to build complex regulatory networks from large expression datasets, but with limited budget on computational resources. In addition, our GPU-centered optimization of adaptive partitioning for Mutual Information (MI) estimation provides lessons that are applicable to other domains.

ARACNe-AP: gene network reverse engineering through adaptive partitioning inference of mutual information.

Summary: The accurate reconstruction of gene regulatory networks from large scale molecular profile datasets represents one of the grand challenges of Systems Biology. The Algorithm for the Reconstruction of Accurate Cellular Networks (ARACNe) represents one of the most effective tools to accomplish this goal. However, the initial Fixed Bandwidth (FB) implementation is both inefficient and unable to deal with sample sets providing largely uneven coverage of the probability density space. Here, we present a completely new implementation of the algorithm, based on an Adaptive Partitioning strategy (AP) for estimating the Mutual Information. The new AP implementation (ARACNe-AP) achieves a dramatic improvement in computational performance (200× on average) over the previous methodology, while preserving the Mutual Information estimator and the Network inference accuracy of the original algorithm. Given that the previous version of ARACNe is extremely demanding, the new version of the algorithm will allow even researchers with modest computational resources to build complex regulatory networks from hundreds of gene expression profiles.Availability and Implementation: A JAVA cross-platform command line executable of ARACNe, together with all source code and a detailed usage guide are freely available on Sourceforge (http://sourceforge.net/projects/aracne-ap). JAVA version 8 or higher is required.Contact: califano@c2b2.columbia.eduSupplementary information: Supplementary data are available at Bioinformatics online.

Combining Expression Data and Knowledge Ontology for Gene Clustering and Network Reconstruction

Many methods have been developed for reverse engineering gene networks from time series expression data. However, when the number of genes and the complexity of regulation increase, it becomes increasingly difficult to infer gene networks. To tackle this scalability problem, this study presents an approach with two phases: gene clustering and network reconstruction. To perform gene clustering, a hybrid method of data and knowledge-based clustering was developed to calculate both data and semantic similarity between genes. In the network reconstruction procedure, a Boolean network model that was inferred from the gene clusters was used to represent the network. A series of experiments were conducted to investigate the effect of the hybrid similarity measure in gene clustering and network reconstruction. The results prove the feasibility and effectiveness of the proposed approach.

Interview with Donald E Ingber.

NanomedicineVol. 9, No. 7 InterviewFree AccessInterview with Donald E IngberDonald E IngberDonald E IngberWyss Institute at Harvard University, 3 Blackfan Circle, Center for Life Sciences Boston, 5th floor, MA 02115, USA E-mail Address: don.ingber@wyss.harvard.eduSearch for more papers by this authorPublished Online:30 Jun 2014https://doi.org/10.2217/nnm.14.31AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinkedInRedditEmail Donald E Ingber, MD, PhD, speaks to Hannah Stanwix, Managing Commissioning Editor: Donald E Ingber is the Founding Director of the Wyss Institute for Biologically Inspired Engineering at Harvard University (MA, USA), Judah Folkman Professor of Vascular Biology at Harvard Medical School and Boston Children's Hospital (MA, USA), and Professor of Bioengineering at the Harvard School of Engineering and Applied Sciences (MA, USA). He received his BA, MA, MPhil, MD and PhD from Yale University (CT, USA). Dr Ingber is a founder of the emerging field of biologically inspired engineering, and at the Wyss Institute, he oversees a multifaceted effort to identify the mechanisms that living organisms use to self-assemble from molecules and cells, and to apply these design principles to develop advanced materials and devices for healthcare and to improve sustainability. In addition, Dr Ingber has made major contributions to mechanobiology, tissue engineering, angiogenesis, systems biology and nanobiotechnology. He was the first to recognize that tensegrity architecture is a fundamental principle that governs molecular self assembly and how living cells self-organize themselves to respond biochemically to mechanical forces. His work on tensegrity has inspired development of DNA self-assembly-based nanotechnologies, and his insights into mechanobiology have led to the creation of shear-activated nanotherapeutics that selectively target clot-busting drugs to sites of vascular occlusion; he also has developed engineered tissues and cancer therapeutics that have entered human clinical trials. Dr Ingber has authored more than 375 publications and 85 patents, and has received numerous honors including the Holst Medal, Pritzker Award from the Biomedical Engineering Society, Rous-Whipple Award from the American Society for Investigative Pathology, Lifetime Achievement Award from the Society of In Vitro Biology and the Department of Defense Breast Cancer Innovator Award. He also serves on the Board of Directors of the National Space Biomedical Research Institute, and is a member of both the American Institute for Medical and Biological Engineering, and the Institute of Medicine of the National Academies.Q Your academic career started with a degree in biophysics and biochemistry. How did you transition from this background into cell biology and bioengineering?I earned both a bachelor's degree and a master's degree in molecular biophysics and biochemistry as a student at Yale University (CT, USA), which actually enabled me to take a broad range of courses in the humanities, art and literature, as well as the sciences. I also studied biology quite extensively, including developmental biology, cancer biology and molecular genetics. But I was studying in the mid 1970s, just before the cloning revolution hit, which meant it was largely classic bacterial genetics. As an undergraduate, I worked with Paul Howard-Flanders on DNA repair at Yale University, and then I won a fellowship to go to London to do research at the Royal Cancer Hospital in Sutton (UK), where I worked with Ken Harrap, who was working on cancer therapeutics at the time. I worked on the effects of steroid analogs on protein phosphorylation in the nucleus, but we were largely taking a cell biology approach, isolating and fractionating cells to look for mechanisms of actions.Q Your early work led to the discovery of a universal set of building rules for organic molecules. What first prompted you to investigate tensegrity?When I came back to Yale University, Ken gave me the name of Alan Sartorelli as someone I may consider working with. He was the Chairman of Pharmacology at Yale University at the time, and later head of the Yale Cancer Center (CT, USA). Sartorelli told me about the projects I could be working on: one that excited me was on cancer metastasis, and on the role of glycosaminoglycans in this process. This is where I first learned how to culture cancer cells and where I first watched them stick and spread on a dish. One of the defining moments in my life was approximately a week after I started in the cancer laboratory, when I was taking an undergraduate sculpture class. It may seem strange that as a molecular biophysics/biochemistry undergraduate student I was taking an art class, but I’ve always been a visual person. I had seen people walking around campus with sculptures that looked like geodesic viruses and so I asked them what course they were studying. They told me it was a course called 'Three-Dimensional Design', and it was a sculpture class. I had tried to get into art classes at Yale University, but it was virtually impossible because there were so many students interested and so few classes. I had given up on it; however, I had a girlfriend at the time who majored in art, and she was a sculptress. The teacher who taught the 3D design class, a man named Erwin Hauer, also taught a human figure class that she was taking. She introduced me to him, and he asked why a scientist would want to take his class. I replied that everything with biology is essentially based on 3D design at the molecular level, such as the DNA helix or actomyosin filament sliding; it's the shape that governs their function – and I got in the class!The first task we were given was to build structures out of sticks and strings, but the sticks could not touch each other; they had to start lifting off the ground as solid structures as a result of being connected by strings. It was a 3-h-long class, and at the beginning nobody had any idea what to do. I think some of the art students must have seen tensegrity sculptures by Kenneth Snelson before and someone eventually worked out how to do it. At the end of the class, Hauer came in and he had a round structure made out of sticks and elastic strings, and as he was talking, he would push it down flat, and then he would let go and it would bounce up in the air. It looked just like the spread cancer cells I was culturing when we clipped their anchors using trypsin and they rounded up and popped off the dish. So I just assumed cells must be built this way.At this time, around 1976, the first papers were coming out about actin and the cytoskeleton, and I was working on a drug called ICF159 in the laboratory that appeared to cause cancer cells to change their shape as part of their action. One day in the laboratory after the sculpture class when we were testing the drug, the cells became extremely flattened and I remember saying to the postdoctoral scholar that “their tensegrity must have changed”. He asked me what I meant and I explained that I was taking a sculpture class on tensegrity structures and Buckminster Fuller (who coined the name 'tensegrity'), and he just cut me off! So I kept quiet and went to the library and I started reading. The more I read, the more I started to see that tensegrity could explain things that could not be explained in biology. I also took a developmental biology course around this time in which we watched videos of cells moving and of embryos developing. These films revealed to me that development is an incredibly physical process, and when I went back to read the early developmental morphologists, they also described virtually all of development in largely mechanical terms. Soon it became clear to me that tensegrity could provide a way to explain how mechanics can influence cell structure and function, and guide development. Soon after I started my MD PhD program at Yale University, I eventually shifted my focus slightly and moved to the laboratory of Jim Jamieson in the Department of Cell Biology where I could study both embryological development and cancer. That is how I really got into cell biology and began to pursue the tensegrity area in the laboratory.Serendipity is an absolutely critical part of science, and certainly, of my life. If I did not have that girlfriend at that time, I would never have been able to take that course and would never have started to think about tensegrity. It is just wild.Q One of your major research focuses is angiogenesis treatment. What led you to initially work in this area?When I was finishing my MD PhD and I was thinking about what I would do after Yale University, I heard that there was a combined research/residency in Boston, but that you needed an advisor to apply. The one person I could think of as a potential advisor was Judah Folkman. Judah had worked on angiogenesis and cancer for many years and everybody knew about his pioneering work in those areas. Most people did not know, however, that in 1978, he had published a major article in the journal Nature that suggested that a cell's shape could regulate its growth. Folkman had actually come to Yale approximately a year earlier to give a lecture, which I learned about at the last minute and attended instead of showing up at another meeting I had scheduled. At the end of his talk, I approached him and handed him a copy of a book chapter on tensegrity and cancer that I originally had intended for the person I was originally supposed to meet with at that time, but I never got a chance to speak to Folkman about what I do or my interests. About a year later, I was finishing my clinical training and Jim Jamieson asked what I wanted to do for my residency. I said a mixed residency at Harvard University would be really good, but the only person I could think of there was Judah Folkman. He smiled and explained that he had recently received a letter from Folkman thanking him for the book chapter on tensegrity and cancer, which had actually confused him. I wrote a letter to Folkman afterwards and he immediately wrote back and I went to visit him. That is how I started to work with him; it was not angiogenesis so much as cell shape.When I visited Folkman, I learned he had discovered that heparin and cortisone inhibited angiogenesis and caused tumors to regress, but he had no idea how it worked. My PhD thesis had revealed that the mechanical properties of the extracellular matrix (ECM) to which cells anchor to in our bodies regulate cell shape and function during tissue development by altering the balance of forces in the tensegrity-stabilized cytoskeleton. Thus, I thought that Folkman's heparin–cortisone combination might work by altering the structure of the capillary ECM. I went to work with him on that, and eventually showed that this was indeed correct, and that we could also inhibit angiogenesis using drugs that specifically target ECM remodeling. We also later confirmed that the ECM regulates angiogenesis and endothelial growth by modulating cell shape and that integrins mediate this mechanotransduction process.Yet another serendipitous thing happened at this time: one day I had a fungal contaminant in my culture dish. Usually this causes the capillary cells to pop off the dish and die, this time when I looked under the microscope, I saw that the fungus did not kill all the cells, it actually induced a gradient of cell rounding. I thought that perhaps the fungus is secreting something that regulates capillary cell shape, which I had just shown regulates cell growth, and hence, that that may secrete an angiogenesis inhibitor. So I decided to culture the fungus, but I did not tell anyone at first because I was an overwhelmed with projects as a postdoctoral fellow and I did not want yet another project. However, eventually, I showed Folkman the fungus, and together we cultured it and identified it as Aspergillus fumigatus fresenius. When we tested it in the standard angiogenesis assays, we confirmed that the product it secreted potently inhibited angiogenesis in various models. A company called Takeda Chemical Industries (now Takeda Pharmaceuticals, Osaka, Japan) thought it was very exciting and funded the project. They identified the specific active molecule as fumagillin, made more than 300 analogs and we both tested them in various model systems. We described the final compound known as TNP-470 (originally called AGM-1470) in a publication in Nature in 1990 [1], and it remains the most generic angiogenesis inhibitor ever found. It was incredibly potent, inhibited virtually any type of growth factor tested and toxicity was low. It worked in small and large animals and so it was moved to humans. In the Phase I trial, there were a couple of complete remissions. Some of the clinicians leading the trials were trying to move the drug to the next phase, but for reasons that are still not clear, at this point TAP Pharmaceuticals (IL, USA; a Takeda Pharmaceuticals partner that was leading the trials in the USA) shut down the trial. The media often describes that the trials were stopped owing to neurotoxicity, but that was not case.Q Do you think nanotechnology could have a role to play here?In fact, owing to general concerns of neurotoxicity, Folkman and Robert D’Amato developed a nanoparticle formulation of TNP-470 called capillostatin that provides the same antiangiogenic activity, while reducing neurotoxicity, and it also can be given orally. But in a more general sense, vascular biology is likely where nanomedicine will have its greatest impact, and it is already beginning to happen. The challenge for most nanotherapeutics is not only injecting nanoparticles, which is relatively easy, but getting the particles through the circulation, across the endothelium, through the tissue and to the right site in the body. I believe that vascular nanomedicine will be the first major success area because injected particles do not have to cross the endothelium or move through dense tissues to exert their actions. It is already happening with imaging agents and the beginnings of targeting moieties and therapeutics. Theranostic nanoparticles could also provide a powerful way to inhibit angiogenesis or other vascular disease, while simultaneously imaging the response to therapy.Q Your recent publications in nanomedicine have been on a variety of topics, including cancer therapy and thrombosis treatment. Could you update us on the areas your group is currently studying?Recently, we have begun to harness our knowledge of mechanics and biology to develop new types of nanotherapeutics. One idea we had was to target the major killers of man, such as atherosclerosis, myocardial infarction, pulmonary embolism and stroke. They all share a common feature, which is that they result from a vascular occlusion, a narrowing of the blood vessel, often from a blood clot. There are great clot-busting drugs out there, such as tissue plasminogen activator (tPA), but only a tiny percentage of patients are eligible to get them owing to the potential of bleeding side effects. Our idea was to use a nanotechnology approach to selectively target and concentrate tPA at clot sites. We started to look at platelets because they target to a narrowing of the blood vessels, and they do this because they are activated by fluid shear stresses that are selectively elevated in these narrowed regions. Inspired by this natural mechanism, we developed a shear-activated drug-delivery system. We used nanoparticles, approximately 180 nm in diameter, that are composed of US FDA-approved biodegradable materials, and we made microscale aggregates of them that are the same size as platelets. We tuned them so if they reach a shear stress that is significantly above normal, they break up into the small nanoparticles, which settle out locally at these sites. In addition, if they are coated with tPA, the released nanoparticles will bind to fibrin clots and remain bound to them until the clots dissolve. When we coated the microaggregates with tPA and injected them into the veins of mice that had experimental pulmonary embolism, we were able to save 85% of the animals' lives, but using only 1/100th the dose of tPA that normally must be administered in solution to produce a similar effect [2]. We are now in the process of commercializing this technology. It is a great example of how harnessing insights from biology and combining them with the power of nanotechnology, you can truly have a unique impact.We have also been exploring the idea that we might be able to reverse cancer rather than kill it by changing the physical and chemical microenvironment of the tumor ECM. We established a mouse transgenic breast cancer model in our laboratory that is driven by a constitutively active oncogene, which is very potent, and all of the animals consistently get breast cancer over the same time course. At approximately 8 weeks old, the mammary glands of the mice are normal, at 12 weeks they are premalignant, at 16 weeks you see small tumors and at 20 weeks you see huge tumors and metastases – 100% of animals follow this pattern. We used this model to start exploring how we can develop differentiation therapies for solid tumors that prevent cancer progression by suppressing growth and inducing differentiation. We collaborated with James Collins who has developed a powerful gene network reverse engineering approach, and we identified genes, in particular HoxA1, which we predicted might drive cancer progression. When we inhibited its expression in vitro using siRNA, it caused the breast cancer cells to stop growing and differentiating quite effectively, but then we had to confront the challenge of delivering this in vivo, and this again brought us back to nanotechnology. We worked with Michael Goldberg, who has developed a lipidoid nanoparticle that can deliver siRNA and provide stable gene suppression for weeks in vivo. Finally, we also wanted to minimize systemic toxicity. Therefore, we wondered if we could deliver the drug right to the mammary duct epithelial cells that are having the problems, rather than administering it systemically. Recently, we showed that we can deliver the HoxA1 siRNA-coated nanoparticles directly into the mammary duct by injecting them through the nipple in transgenic mice. Injection of this nanotherapeutic prevented progression to cancer in 75% of animals [3], which is really mind-boggling given that all of the cells still express the active oncogene at high levels. We hope in the future to leverage this type of approach to develop nanomaterials that you might sprinkle onto a primary tumor site after a surgical excision to induce any remaining cells to differentiate and prevent the tumor from regrowing. We are also exploring injectable nanoparticles that could travel to distant metastatic sites to prevent a dormant tumor from becoming cancerous.Q Another of your areas of interest is the development of ‘organ-on-a-chip’ systems. Could you explain to us the rationale behind this work?We have a major effort in Biomimetic Microsystems, or what have come to be known as human ‘organs-on-chips’. We use microfabrication techniques to make microfluidic devices with hollow channels that we line with living human cells. We have made a lung-on-a-chip with human lung epithelial cells on top of a porous membrane coated with ECM, and human lung capillary cells on the bottom of the same membrane, which is stretched across the microfluidic channel. On the top of the lung cells, we have air, and we flow medium over the tops of the capillary cells, so we have essentially rebuilt the alveolar–capillary interface of the lung air sac. In addition, the walls of the device are flexible, and we apply cyclic suction to side chambers that cause the tissue–tissue interface to rhythmically stretch and relax, hence faithfully mimicking the breathing motions of the lung. We demonstrated that using this lung-on-a-chip, we can measure drug efficacies, detect toxicities and model human diseases, such as pulmonary edema [4,5]. We also used these devices to study nanotoxicology in the lung. For example, we showed that nanoparticle simulants of airborne particulates only induce injury and inflammation in the lung-on-a-chip when exposed to physiological breathing motions. We also discovered that breathing increases the absorption of nanoparticles from the air space into the vascular channel by almost tenfold, and that these mechanical motions are critical for induction of pulmonary edema by the cancer drug IL-2. Moreover, we confirmed these predictions in animal models. We are now developing more than ten different human organ chips, and linking them to together to form a human ‘body-on-a-chip’ with the goal of replacing animal models over time for drug development, and testing of chemicals and cosmetics, among others.Q You are the Founding Director of Wyss Institute for Biologically Inspired Engineering at Harvard University. What are the main research projects currently underway at the Institute?The institute is extremely broad, with both medical and nonmedical projects ongoing. It was based on the idea that engineering has transformed medicine over the past 50 years, but we feel that we have learned so much from biology that it is now time to leverage biological principles to develop new engineering innovations, particularly in the areas of materials, devices and nanotechnologies. One thing that we really focus on is translation; getting fundamental discoveries out and into the clinic and into the market place. We have business development people and entrepreneurs-in-residence working side-by-side with students, fellows, staff and faculty. We focus on intellectual property generation and protection, and we employ over 40 expert scientists and engineers who bring extensive industrial experience in product development into our effort. It is a new model for academia.Looking at some specific projects, David Mooney leads a platform focused on programmable nanomaterials. The lead project is a cancer vaccine using a nanofibrous polymer scaffold. They have found their approach causes complete regression in approximately 50–90% of tumors in animal models. The results were so exciting in melanoma that they quickly obtained FDA approval to start a human clinical trial at the Dana Farber Cancer Institute (MA, USA). George Church has a platform in synthetic biology, and along with James Collins and Pam Silver, they are engineering microbes and developing cheaper and faster DNA sequencing and DNA synthesis approaches. William Shih, Peng Yin and George Church are also doing work with programmable DNA-based nanomaterials, using DNA origami-based self-assembly approaches, which has also attracted great attention recently. Interestingly, William and I collaborated to fabricate a circular DNA molecule that when combined with DNA 'staples' self assembles into a 3D prestressed tensegrity structure, just like one of the stick-and-string models I built in my undergraduate sculpture class, but on the 10-nm scale.Q Of your more than 375 publications, if you had to choose one as your favorite, which would it be and why?My favorite publication is the one that people have never read: it is a theoretical book chapter entitled, “Cells as tensegrity structures: architectural regulation of histodifferentiation by physical forces transduced over basement membrane”, which I published in 1985 [6] that essentially lays out all of my strange ideas relating to tensegrity, mechanobiology, development and cancer that I pursued for the rest of my career, and eventually confirmed experimentally. It was also literally the last chapter of my PhD dissertation. In terms of high-impact papers, my Science paper in 1993 showing that integrins mediate mechanotransduction, and also that cells behave like transegrity structures, is very important to me [7]. I published another article in Science with George Whitesides in 1994 called “Engineering cell shape and function” [8] that was the first to use computer micromanufacturing methods combined with a nanotechnology-based molecular self assembled monolayer technique to fabricate culture substrates that control cell function by dictating their shape and position on the micrometer scale. It was a very high-impact paper with thousands of citations, but not as high impact as a later paper that we also published in Science in 1997 called, “Geometric control of cell life and death”. In this article, we unequivocally showed that cell shape distortion is a key regulator of cell fate, and that cells can be switched between growth and apoptosis by altering cell spreading independently of ECM adhesion or growth factors. This paper has now been cited approximately 3000 times [9].Q Finally, if you did not work in science, what would you do?That is a hard one! I have done some film work and writing in the past; and I love to write, so it be writing in some way. It would probably be something such as a writer/director for television, film or some other visual communications medium. It combines the visual aspect with writing, communication, art and humor.AcknowledgementsThe author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert t­estimony, grants or patents received or pending, or royalties.No writing assistance was utilized in the production of this manuscript.References1 Ingber DE, Fujita T, Kishimoto S et al. Synthetic analogues of fumagillin which inhibit angiogenesis and suppress tumour growth. Nature 348, 555–557 (1990).Crossref, Medline, CAS, Google Scholar2 Korin N, Kanapathipillai M, Matthew BD et al. Shear-activated nanotherapeutics for drug targeting to obstructed blood vessels. Science 337, 738–742 (2012).Crossref, Medline, CAS, Google Scholar3 Brock A, Krause S, Li H et al. Silencing HoxA1 by intraductal injection of siRNA lipidoid nanoparticles prevents mammary tumor progression in mice. Sci. Trans. Med. 6(217), 6:217ra2 (2014).Crossref, Medline, Google Scholar4 Huh D, Matthews BD, Mammoto A, Montoya-Zavala M, Hsin HY, Ingber DE. Reconstituting organ-level lung functions on a chip. Science 328, 1662–1668 (2010).Crossref, Medline, CAS, Google Scholar5 Huh D, Leslie DC, Matthews BD et al. A human disease model of drug toxicity-induced pulmonary edema in a lung-on-a-chip microdevice. Sci. Trans. Med 4, 159ra147 (2012).Crossref, Medline, Google Scholar6 Ingber DE, Jamieson JD. Cells as tensegrity structures: architectural regulation of histodifferentiation by physical forces tranduced over basement membrane. In: Gene Expression During Normal and Malignant Differentiation. Andersson LC, Gahmberg CG, Ekblom P (Eds). Academic Press, FL, USA 13–32 (1985).Google Scholar7 Wang N, Butler JP, Ingber D. Mechanotransduction across the cell surface and through the cytoskeleton. Science 260(5111), 1124–1127 (1993).Crossref, Medline, CAS, Google Scholar8 Singhvi R, Kumar A, Lopez G et al. Engineering cell shape and function. Science 264, 696–698 (1994).Crossref, Medline, CAS, Google Scholar9 Chen CS, Mrksich M, Huang S, Whitesides GM, Ingber DE. Geometric control of cell life and death. Science 276(5317), 1425–1428 (1997).Crossref, Medline, CAS, Google ScholarFiguresReferencesRelatedDetailsCited ByProfessor Lev Beloussov and the birth of morphomechanicsBiosystems, Vol. 173The Theory of Tensegrity and Spatial Organization of Living Matter3 April 2018 | Russian Journal of Developmental Biology, Vol. 49, No. 2A Re-Examination of Free SpeechSSRN Electronic Journal Vol. 9, No. 7 Follow us on social media for the latest updates Metrics History Published online 30 June 2014 Published in print May 2014 Information© Future Medicine LtdAcknowledgementsThe author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert t­estimony, grants or patents received or pending, or royalties.No writing assistance was utilized in the production of this manuscript.PDF download

Comparative study of RNA-seq- and Microarray-derived coexpression networks in Arabidopsis thaliana

Coexpression networks are data-derived representations of genes behaving in a similar way across tissues and experimental conditions. They have been used for hypothesis generation and guilt-by-association approaches for inferring functions of previously unknown genes. So far, the main platform for expression data has been DNA microarrays; however, the recent development of RNA-seq allows for higher accuracy and coverage of transcript populations. It is therefore important to assess the potential for biological investigation of coexpression networks derived from this novel technique in a condition-independent dataset. We collected 65 publicly available Illumina RNA-seq high quality Arabidopsis thaliana samples and generated Pearson correlation coexpression networks. These networks were then compared with those derived from analogous microarray data. We show how Variance-Stabilizing Transformed (VST) RNA-seq data samples are the most similar to microarray ones, with respect to inter-sample variation, correlation coefficient distribution and network topological architecture. Microarray networks show a slightly higher score in biology-derived quality assessments such as overlap with the known protein-protein interaction network and edge ontological agreement. Different coexpression network centralities are investigated; in particular, we show how betweenness centrality is generally a positive marker for essential genes in A.thaliana, regardless of the platform originating the data. In the end, we focus on a specific gene network case, showing that although microarray data seem more suited for gene network reverse engineering, RNA-seq offers the great advantage of extending coexpression analyses to the entire transcriptome.

Profile of James J. Collins

BioTechniquesVol. 53, No. 4 ScientistsOpen AccessProfile of James J. CollinsKristie NyboKristie NyboSearch for more papers by this authorPublished Online:3 Apr 2018https://doi.org/10.2144/000113932AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinkedInRedditEmail Jim Collins' work on programmable cells and his contributions to synthetic biology caught our attention. Curious to know more, BioTechniques contacted him to find out about the ambition, character, and motivation that led to his success.Forward engineeringHow did you first become interested in bioengineering and synthetic biology?I became interested in bioengineering at a young age when both of my grandfathers became disabled: one lost his vision and the other had a series of strokes and became hemiplegic. My father did engineering work for NASA and the military, so I often heard about marvelous technologies developed to shoot things into the sky, but I saw very little being developed to help my grandfathers regain function. This motivated me to harness engineering and science insights to enhance the lives of those affected by disease or injury.When I joined the faculty at Boston University, my lab focused on whole body dynamics and biomechanics including how humans walk, run, and balance. In that context, we developed a number of interesting devices to enhance sensory and motor function. However, over the last several years, I have made a pretty dramatic shift into systems biology and synthetic biology, which started when my colleagues Charles Cantor and Charles DeLisi encouraged me to think about reverse engineering gene networks using data from the Human Genome Project. We set about applying engineering control theory techniques and non-linear dynamics principles at the molecular level, but quickly recognized there was insufficient expression data at that time to reverse-engineer natural networks. Molecular biology was primed for forward engineering approaches though.Up to this point, what has been your most significant contribution to your field?We helped to launch the field of synthetic biology by creating a bistable genetic toggle switch in Escherichia coli, one of the first synthetic gene networks. We used engineering principles to create a circuit schematic, identified the appropriate components including promoters, proteins, and bits of DNA, and assembled them with techniques developed for recombinant DNA and genetic engineering applications. We designed the switch as a mutually inhibitory network that included two repressor genes driven by constitutive promoters; the protein product from gene 1 shuts down the promoter for gene 2, while the protein product from gene 2 binds the promoter for gene 1, shutting it down. So in principle, this system may exist in two states: gene 1 is on and gene 2 is off or vice versa. We can switch between states by delivering an inducer for the currently active gene, using this control circuit to regulate, in a bistable fashion, the expression of up to a relatively large number of proteins.What is the application for these synthetic circuits?Basically, they are a means to endow programmable cells with a simple form of memory. For example, we have used toggle switches to create whole cell biosensors that detect heavy metals, chemicals, or even pathogens in their environments, designed genetic switchboards for metabolic engineering applications, and we have plans to use them in the context of cell and gene therapy. The advantage of toggle switches in these situations is that switching between the on and off states requires only transient delivery of a chemical stimulus rather than continuous application.What are you working on now?We are very excited now about designing programmable cell sentinels that can be used to sense and respond to viral or bacterial pathogens. We are exploring the possibility of modifying the human microbiome with engineered cells capable of detecting and responding to cholera infections. In the absence of an infection, the bacteria simply live and divide in the host; but when small molecules produced during a cholera infection are detected by the circuit, it switches on and releases an antimicrobial peptide to counter the invading cholera.We are also looking into platforms for treating tuberculosis (TB) by delivering synthetic circuitry into macrophage or granuloma cells. Rather than engineering bacteria for this application, we are planning to create synthetic vesicles, engineered liposomes for example, that can be administered to a patient by nebulizer following identification of a TB infection.Do you have any additional projects outside the main focus of your lab?By combining our engineering approaches with George Daley's group's outstanding expertise in stem cells, we are trying to gain additional insight into reprogramming and differentiation of embryonic stem (ES) cells or induced pluripotent stem (iPS) cells for biomedical applications. We use systems and synthetic biology methods to study the regulatory networks underlying reprogramming to identify the components needed to produce specific cell and tissue types of interest from ES or iPS cells for regenerative medicine.What is the biggest challenge in synthetic biology today?The biggest challenge for the field right now is developing the ability to predictably design integrated circuits, components, and pathways that function as we desire. That challenge will require an expanded understanding of the underlying biology of these components and how they function, the development of better models for interactions that are currently unknown or unaccounted for in our modeling, and the development of expanded libraries of well characterized parts that can be harnessed for such design purposes.FiguresReferencesRelatedDetails Vol. 53, No. 4 Follow us on social media for the latest updates Metrics History Published online 3 April 2018 Published in print October 2012 Information© 2012 Author(s)PDF download

An S-System Parameter Estimation Method (SPEM) for Biological Networks

Advances in experimental biology, coupled with advances in computational power, bring new challenges to the interdisciplinary field of computational biology. One such broad challenge lies in the reverse engineering of gene networks, and goes from determining the structure of static networks, to reconstructing the dynamics of interactions from time series data. Here, we focus our attention on the latter area, and in particular, on parameterizing a dynamic network of oriented interactions between genes. By basing the parameterizing approach on a known power-law relationship model between connected genes (S-system), we are able to account for non-linearity in the network, without compromising the ability to analyze network characteristics. In this article, we introduce the S-System Parameter Estimation Method (SPEM). SPEM, a freely available R software package (http://www.picb.ac.cn/ClinicalGenomicNTW/temp3.html), takes gene expression data in time series and returns the network of interactions as a set of differential equations. The methods, which are presented and tested here, are shown to provide accurate results not only on synthetic data, but more importantly on real and therefore noisy by nature, biological data. In summary, SPEM shows high sensitivity and positive predicted values, as well as free availability and expansibility (because based on open source software). We expect these characteristics to make it a useful and broadly applicable software in the challenging reconstruction of dynamic gene networks.

Insights gained from the reverse engineering of gene networks in keloid fibroblasts

BackgroundKeloids are protrusive claw-like scars that have a propensity to recur even after surgery, and its molecular etiology remains elusive. The goal of reverse engineering is to infer gene networks from observational data, thus providing insight into the inner workings of a cell. However, most attempts at modeling biological networks have been done using simulated data. This study aims to highlight some of the issues involved in working with experimental data, and at the same time gain some insights into the transcriptional regulatory mechanism present in keloid fibroblasts.MethodsMicroarray data from our previous study was combined with microarray data obtained from the literature as well as new microarray data generated by our group. For the physical approach, we used the fREDUCE algorithm for correlating expression values to binding motifs. For the influence approach, we compared the Bayesian algorithm BANJO with the information theoretic method ARACNE in terms of performance in recovering known influence networks obtained from the KEGG database. In addition, we also compared the performance of different normalization methods as well as different types of gene networks.ResultsUsing the physical approach, we found consensus sequences that were active in the keloid condition, as well as some sequences that were responsive to steroids, a commonly used treatment for keloids. From the influence approach, we found that BANJO was better at recovering the gene networks compared to ARACNE and that transcriptional networks were better suited for network recovery compared to cytokine-receptor interaction networks and intracellular signaling networks. We also found that the NFKB transcriptional network that was inferred from normal fibroblast data was more accurate compared to that inferred from keloid data, suggesting a more robust network in the keloid condition.ConclusionsConsensus sequences that were found from this study are possible transcription factor binding sites and could be explored for developing future keloid treatments or for improving the efficacy of current steroid treatments. We also found that the combination of the Bayesian algorithm, RMA normalization and transcriptional networks gave the best reconstruction results and this could serve as a guide for future influence approaches dealing with experimental data.

A Swarm Intelligence Framework for Reconstructing Gene Networks: Searching for Biologically Plausible Architectures

In this paper, we investigate the problem of reverse engineering the topology of gene regulatory networks from temporal gene expression data. We adopt a computational intelligence approach comprising swarm intelligence techniques, namely particle swarm optimization (PSO) and ant colony optimization (ACO). In addition, the recurrent neural network (RNN) formalism is employed for modeling the dynamical behavior of gene regulatory systems. More specifically, ACO is used for searching the discrete space of network architectures and PSO for searching the corresponding continuous space of RNN model parameters. We propose a novel solution construction process in the context of ACO for generating biologically plausible candidate architectures. The objective is to concentrate the search effort into areas of the structure space that contain architectures which are feasible in terms of their topological resemblance to real-world networks. The proposed framework is initially applied to the reconstruction of a small artificial network that has previously been studied in the context of gene network reverse engineering. Subsequently, we consider an artificial data set with added noise for reconstructing a subnetwork of the genetic interaction network of S. cerevisiae (yeast). Finally, the framework is applied to a real-world data set for reverse engineering the SOS response system of the bacterium Escherichia coli. Results demonstrate the relative advantage of utilizing problem-specific knowledge regarding biologically plausible structural properties of gene networks over conducting a problem-agnostic search in the vast space of network architectures.

From Knockouts to Networks: Establishing Direct Cause-Effect Relationships through Graph Analysis

BackgroundReverse-engineering gene networks from expression profiles is a difficult problem for which a multitude of techniques have been developed over the last decade. The yearly organized DREAM challenges allow for a fair evaluation and unbiased comparison of these methods.ResultsWe propose an inference algorithm that combines confidence matrices, computed as the standard scores from single-gene knockout data, with the down-ranking of feed-forward edges. Substantial improvements on the predictions can be obtained after the execution of this second step.ConclusionsOur algorithm was awarded the best overall performance at the DREAM4 In Silico 100-gene network sub-challenge, proving to be effective in inferring medium-size gene regulatory networks. This success demonstrates once again the decisive importance of gene expression data obtained after systematic gene perturbations and highlights the usefulness of graph analysis to increase the reliability of inference.

TimeDelay-ARACNE: Reverse engineering of gene networks from time-course data by an information theoretic approach.

BackgroundOne of main aims of Molecular Biology is the gain of knowledge about how molecular components interact each other and to understand gene function regulations. Using microarray technology, it is possible to extract measurements of thousands of genes into a single analysis step having a picture of the cell gene expression. Several methods have been developed to infer gene networks from steady-state data, much less literature is produced about time-course data, so the development of algorithms to infer gene networks from time-series measurements is a current challenge into bioinformatics research area. In order to detect dependencies between genes at different time delays, we propose an approach to infer gene regulatory networks from time-series measurements starting from a well known algorithm based on information theory.ResultsIn this paper we show how the ARACNE (Algorithm for the Reconstruction of Accurate Cellular Networks) algorithm can be used for gene regulatory network inference in the case of time-course expression profiles. The resulting method is called TimeDelay-ARACNE. It just tries to extract dependencies between two genes at different time delays, providing a measure of these dependencies in terms of mutual information. The basic idea of the proposed algorithm is to detect time-delayed dependencies between the expression profiles by assuming as underlying probabilistic model a stationary Markov Random Field. Less informative dependencies are filtered out using an auto calculated threshold, retaining most reliable connections. TimeDelay-ARACNE can infer small local networks of time regulated gene-gene interactions detecting their versus and also discovering cyclic interactions also when only a medium-small number of measurements are available. We test the algorithm both on synthetic networks and on microarray expression profiles. Microarray measurements concern S. cerevisiae cell cycle, E. coli SOS pathways and a recently developed network for in vivo assessment of reverse engineering algorithms. Our results are compared with ARACNE itself and with the ones of two previously published algorithms: Dynamic Bayesian Networks and systems of ODEs, showing that TimeDelay-ARACNE has good accuracy, recall and F-score for the network reconstruction task.ConclusionsHere we report the adaptation of the ARACNE algorithm to infer gene regulatory networks from time-course data, so that, the resulting network is represented as a directed graph. The proposed algorithm is expected to be useful in reconstruction of small biological directed networks from time course data.

Challenges in Reverse Engineering of Gene Networks from Algebraic Perspectives

Reverse engineering of complex gene networks is a significant step towards understanding various biological processes. In this work, a novel algorithm for reverse engineering gene networks on a genome-wide scale using a noisy gene expression data is proposed. Under the proposed scheme, the challenges in reverse engineering of gene networks from these gene expression datasets are highlighted. Also, the parameter space describing gene interaction is partitioned into estimable and inestimable linear subspaces. The estimable linear subspace is obtained by using principal components analysis (PCA), the Akaike information criterion (AIC), and jackknifing. Furthermore, the approach is tested and validated using a simulated gene network model.

Reverse engineering and verification of gene networks: Principles, assumptions, and limitations of present methods and future perspectives

Reverse engineering of gene networks aims at revealing the structure of the gene regulation network in a biological system by reasoning backward directly from experimental data. Many methods have recently been proposed for reverse engineering of gene networks by using gene transcript expression data measured by microarray. Whereas the potentials of the methods have been well demonstrated, the assumptions and limitations behind them are often not clearly stated or not well understood. In this review, we first briefly explain the principles of the major methods, identify the assumptions behind them and pinpoint the limitations and possible pitfalls in applying them to real biological questions. With regard to applications, we then discuss challenges in the experimental verification of gene networks generated from reverse engineering methods. We further propose an optimal experimental design for allocating sampling schedule and possible strategies for reducing the limitations of some of the current reverse engineering methods. Finally, we examine the perspectives for the development of reverse engineering and urge the need to move from revealing network structure to the dynamics of biological systems.

Reverse Engineering of Gene Networks with LASSO and Nonlinear Basis Functions

The quest to determine cause from effect is often referred to as reverse engineering in the context of cellular networks. Here we propose and evaluate an algorithm for reverse engineering a gene regulatory network from time-series and steady-state data. Our algorithmic pipeline, which is rather standard in its parts but not in its integrative composition, combines ordinary differential equations, parameter estimations by least angle regression, and cross-validation procedures for determining the in-degrees and selection of nonlinear transfer functions. The result of the algorithm is a complete directed network, in which each edge has been assigned a score from a bootstrap procedure. To evaluate the performance, we submitted the outcome of the algorithm to the reverse engineering assessment competition DREAM2, where we used the data corresponding to the InSilico1 and InSilico2 networks as input. Our algorithm outperformed all other algorithms when inferring one of the directed gene-to-gene networks.

Replaying the Evolutionary Tape: Biomimetic Reverse Engineering of Gene Networks

In this paper, we suggest a new approach for reverse engineering gene regulatory networks, which consists of using a reconstruction process that is similar to the evolutionary process that created these networks. The aim is to integrate prior knowledge into the reverse-engineering procedure, thus biasing the search toward biologically plausible solutions. To this end, we propose an evolutionary method that abstracts and mimics the natural evolution of gene regulatory networks. Our method can be used with a wide range of nonlinear dynamical models. This allows us to explore novel model types such as the log-sigmoid model introduced here. We apply the biomimetic method to a gold-standard dataset from an in vivo gene network. The obtained results won a reverse engineering competition of the second DREAM conference (Dialogue on Reverse Engineering Assessments and Methods 2007, New York, NY).

Generating Realistic In Silico Gene Networks for Performance Assessment of Reverse Engineering Methods

Reverse engineering methods are typically first tested on simulated data from in silico networks, for systematic and efficient performance assessment, before an application to real biological networks. In this paper, we present a method for generating biologically plausible in silico networks, which allow realistic performance assessment of network inference algorithms. Instead of using random graph models, which are known to only partly capture the structural properties of biological networks, we generate network structures by extracting modules from known biological interaction networks. Using the yeast transcriptional regulatory network as a test case, we show that extracted modules have a biologically plausible connectivity because they preserve functional and structural properties of the original network. Our method was selected to generate the "gold standard" networks for the gene network reverse engineering challenge of the third DREAM conference (Dialogue on Reverse Engineering Assessment and Methods 2008, Cambridge, MA).

Comparative analysis of microarray normalization procedures: effects on reverse engineering gene networks

An increasingly common application of gene expression profile data is the reverse engineering of cellular networks. However, common procedures to normalize expression profiles generated using the Affymetrix GeneChips technology were originally developed for a rather different purpose, namely the accurate measure of differential gene expression between two or more phenotypes. As a result, current evaluation strategies lack comprehensive metrics to assess the suitability of available normalization procedures for reverse engineering and, in general, for measuring correlation between the expression profiles of a gene pair. We benchmark four commonly used normalization procedures (MAS5, RMA, GCRMA and Li-Wong) in the context of established algorithms for the reverse engineering of protein-protein and protein-DNA interactions. Replicate sample, randomized and human B-cell data sets are used as an input. Surprisingly, our study suggests that MAS5 provides the most faithful cellular network reconstruction. Furthermore, we identify a crucial step in GCRMA responsible for introducing severe artifacts in the data leading to a systematic overestimate of pairwise correlation. This has key implications not only for reverse engineering but also for other methods, such as hierarchical clustering, relying on accurate measurements of pairwise expression profile correlation. We propose an alternative implementation to eliminate such side effect.