Polly Fordyce is an Assistant Professor of Genetics and Bioengineering and fellow of the ChEM-H Institute at Stanford, where her laboratory focuses on developing and applying new microfluidic platforms for quantitative, high-throughput biophysics and biochemistry and single-cell genomics. She graduated from the University of Colorado at Boulder with undergraduate degrees in physics and biology before moving to Stanford University, where she earned a Ph.D. in physics for work with Professor Steve Block developing instrumentation and assays for single-molecule studies of kinesin motor proteins. For her postdoctoral research, she worked with Professor Joe DeRisi to develop a new microfluidic platform for understanding how transcription factors recognize and bind their DNA targets as well as a new technology for bead-based multiplexing. She is the recipient of a number of awards, including an NIH New Innovator Award and an Alfred P. Sloan Foundation Research Fellowship, and is a Chan Zuckerberg Biohub Investigator.
Investigator, Chan Zuckerberg Biohub (2017 - 2022)
Honors & Awards
Breakthrough Science Initiative Award, Ono Pharma Foundation (2019-2022)
Investigator, Chan Zuckerberg Biohub (2017-2022)
Alfred P. Sloan Foundation Research Fellow, Alfred P. Sloane Foundarion (2017-2019)
New Innovator Award (DP2), NIH (2016-2021)
Scialog Fellow, Gordon & Betty Moore Foundation (2016-2017)
Pathway to Independence Award (K99), NIH (2012-2014)
Helen Hay Whitney Postdoctoral Fellowship, Helen Hay Whitney Foundation (2008-2011)
G. J. Lieberman Fellow, Stanford University (2003-2004)
Graduate Research Fellow, National Science Foundation (2002-2005)
Boards, Advisory Committees, Professional Organizations
Advisory Board Member, Cell Systems (2020 - Present)
Postdoctoral Fellow, University of California San Francisco, Biophysics (2014)
Ph.D., Stanford University, Physics (2007)
B.A., University of Colorado at Boulder, Physics, Biology (2000)
Yinnian Feng, Adam White, Jamin Hein, and Polly Fordyce. "United States Patent 63/037,804 Methods, devices, and compositions related to polymeric microbeads", Chan Zuckerberg Biohub, Jun 11, 2020
Kara Brower, Sandy Klemm, William Greenleaf, Polly Fordyce. "United States Patent 62/693,800 Method to perform high-throughput single cell genomic and phenotypic analyses", Chan Zuckerberg Biohub, Jul 1, 2019
Adam White, Huy Nguyen, Brian Yu, Tyler Shimko, Polly Fordyce, Nadya Andini, Sam Yang. "United States Patent 62/853,494 Method for multiplexed detection of nucleic acids using spectrally encoded beads", Chan Zuckerberg Biohub, May 28, 2019
Kara Brower, Alex Sockell, Adam White, Polly Fordyce. "United States Patent 62/853,627 Multi-parameter single-cell analysis using spectrally encoded microbeads", Chan Zuckerberg Biohub, May 28, 2019
Brian Baxter, Joe DeRisi, Polly Fordyce, Rachel Gerver, Rafael Gomez-Sjoberg, Kurt Thorn. "United States Patent 61/692,816. Spectrally encoded microbeads and methods and devices for making and using same", University of California San Francisco, Aug 23, 2013
Current Research and Scholarly Interests
Cellular function and organismal homeostasis are governed by molecular interactions. Protein-DNA binding interactions are essential for regulating gene transcription and translation, dense networks of protein-protein and protein-peptide interactions further regulate cellular function, and enzymes make possible all of the chemical transformations essential to metabolism and signaling. Our goal is to understand, and eventually engineer, these complex processes by building and testing biophysical models of how the molecules that drive these processes work. To do so, an essential first step is to obtain the necessary quantitative measurements of the fundamental kinetic and thermodynamic constants of these molecular interactions and catalytic processes—the “universal language” needed to describe and ultimately predict function. In our lab, we use microfluidics and extensive hardware automation to perform these quantitative measurements at an unprecedented scale via 3 main platforms:
1. Array-based multiplexing experiments (MITOMI and HT-MEK) employ microfluidic devices containing 1,568 valved reaction chambers aligned to printed DNA arrays. We are currently using these devices to better understand how transcription factors find and bind their genomic targets to regulate gene expression, as well as to understand how enzymes achieve their extraordinary catalytic efficiency and substrate specificity.
2. MRBLEs (Microspheres with Ratiometric Barcode Lanthanide Encoding) rely on spectral multiplexing to track analytes throughout an experiment. We can create microspheres containing > 1,000 distinct ratios of lanthanide nanophosphors that can be uniquely identified via imaging alone, and are now using these MRBLEs in a variety of downstream assays.
3. Dropception is a microfluidic platform for creating double emulsion (water-in-oil-in-water) droplets that can be sorted in high-throughput using standard flow cytometers (FACS machines). We recently demonstrated the ability to generate and sort double emulsion droplets without breakage, isolate individual rare droplets of interest in wells of a multiwell plate, and recover all encapsulated nucleic acids, enabling a wide range of novel single-cell multi-omic techniques.
- Bioengineering Departmental Research Colloquium
BIOE 393 (Aut)
Independent Studies (11)
- Bioengineering Problems and Experimental Investigation
BIOE 191 (Win, Sum)
- Directed Investigation
BIOE 392 (Aut, Win, Spr, Sum)
- Directed Reading in Biophysics
BIOPHYS 399 (Win, Spr)
- Directed Reading in Genetics
GENE 299 (Aut, Win, Spr)
- Directed Studies in Applied Physics
APPPHYS 290 (Sum)
- Directed Study
BIOE 391 (Aut)
- Graduate Research
BIOPHYS 300 (Aut, Win, Spr, Sum)
- Graduate Research
GENE 399 (Aut, Win, Spr, Sum)
- Medical Scholars Research
GENE 370 (Aut, Win, Spr)
- Supervised Study
GENE 260 (Aut, Win, Spr, Sum)
- Undergraduate Research
GENE 199 (Aut, Win, Spr)
- Bioengineering Problems and Experimental Investigation
- Prior Year Courses
Doctoral Dissertation Reader (AC)
Namrata Anand, Katie Antilla, Abhimanyu Banerjee, Binbin Chen, Robert Coukos, Rebecca Culver, Shreya Deshmukh, Anthony Ho, Athena Ierokomos, Margarita Khariton, Chelsea Longwell, Elisabeth Meyer, Margaux Pinney, Abhiram Rao, Suhas Rao, Neeraja Ravi, Avanti Shrikumar, Chris Still, Alec Tarashansky, Terence Theisen, Nicole Xu
Postdoctoral Faculty Sponsor
Yinnian Feng, Samuel Thompson, Adam White, Zheng Zuo
Doctoral Dissertation Advisor (AC)
Arjun Aditham, Eliel Akinbami, Beatriz Atsavapranee, Kara Brower, Renee Hastings, Scott Longwell, Peter Suzuki
Chelsea Longwell, Avanti Shrikumar
Doctoral Dissertation Co-Advisor (AC)
Nicole DelRosso, Amalia Hadjitheodorou, Daniel Mokhtari, Alexandra Sockell
Beatriz Atsavapranee, Manish Ayushman, Sam Bray, Caitlyn Miller, Neeraja Ravi, Avin Veerakumar, Cassandra Villicana, Alexa Wnorowski
Double emulsion flow cytometry with high-throughput single droplet isolation and nucleic acid recovery.
Lab on a chip
Droplet microfluidics has made large impacts in diverse areas such as enzyme evolution, chemical product screening, polymer engineering, and single-cell analysis. However, while droplet reactions have become increasingly sophisticated, phenotyping droplets by a fluorescent signal and sorting them to isolate individual variants-of-interest at high-throughput remains challenging. Here, we present sdDE-FACS (s[combining low line]ingle d[combining low line]roplet D[combining low line]ouble E[combining low line]mulsion-FACS), a new method that uses a standard flow cytometer to phenotype, select, and isolate individual double emulsion droplets of interest. Using a 130 mum nozzle at high sort frequency (12-14 kHz), we demonstrate detection of droplet fluorescence signals with a dynamic range spanning 5 orders of magnitude and robust post-sort recovery of intact double emulsion (DE) droplets using 2 commercially-available FACS instruments. We report the first demonstration of single double emulsion droplet isolation with post-sort recovery efficiencies >70%, equivalent to the capabilities of single-cell FACS. Finally, we establish complete downstream recovery of nucleic acids from single, sorted double emulsion droplets via qPCR with little to no cross-contamination. sdDE-FACS marries the full power of droplet microfluidics with flow cytometry to enable a variety of new droplet assays, including rare variant isolation and multiparameter single-cell analysis.
View details for DOI 10.1039/d0lc00261e
View details for PubMedID 32417874
A High-Throughput Assay Platform for Next-Generation Mechanistic Enzymology and Applications
CELL PRESS. 2020: 535A
View details for Web of Science ID 000513023203420
A High-Throughput Platform Characterizes Functional Effects of Transcription Factor Mutations
CELL PRESS. 2020: 74A–75A
View details for Web of Science ID 000513023200373
DeCoDe: degenerate codon design for complete protein-coding DNA libraries.
Bioinformatics (Oxford, England)
High-throughput protein screening is a critical technique for dissecting and designing protein function. Libraries for these assays can be created through a number of means, including targeted or random mutagenesis of a template protein sequence or direct DNA synthesis. However, mutagenic library construction methods often yield vastly more non-functional than functional variants and, despite advances in large-scale DNA synthesis, individual synthesis of each desired DNA template is often prohibitively expensive. Consequently, many protein screening libraries rely on the use of degenerate codons (DCs), mixtures of DNA bases incorporated at specific positions during DNA synthesis, to generate highly diverse protein variant pools from only a few low-cost synthesis reactions. However, selecting DCs for sets of sequences that covary at multiple positions dramatically increases the difficulty of designing a DC library and leads to the creation of many undesired variants that can quickly outstrip screening capacity.We introduce a novel algorithm for total DC library optimization, DeCoDe, based on integer linear programming. DeCoDe significantly outperforms state-of-the-art DC optimization algorithms and scales well to more than a hundred proteins sharing complex patterns of covariation (e.g., the lab-derived avGFP lineage). Moreover, DeCoDe is, to our knowledge, the first DC design algorithm with the capability to encode mixed-length protein libraries. We anticipate DeCoDe to be broadly useful for a variety of library generation problems, ranging from protein engineering attempts that leverage mutual information to the reconstruction of ancestral protein states.github.com/OrensteinLab/DeCoDe.Supplementary data are available at Bioinformatics online.
View details for DOI 10.1093/bioinformatics/btaa162
View details for PubMedID 32176271
Quantitative mapping of protein-peptide affinity landscapes using spectrally encoded beads.
Transient, regulated binding of globular protein domains to Short Linear Motifs (SLiMs) in disordered regions of other proteins drives cellular signaling. Mapping the energy landscapes of these interactions is essential for deciphering and perturbing signaling networks but is challenging due to their weak affinities. We present a powerful technology (MRBLE-pep) that simultaneously quantifies protein binding to a library of peptides directly synthesized on beads containing unique spectral codes. Using MRBLE-pep, we systematically probe binding of human calcineurin (CN), a conserved protein phosphatase essential for the immune response and target of immunosuppressants, to the PxIxIT SLiM. We discover that flanking residues and post-translational modifications critically contribute to PxIxIT-CN affinity and identify CN-binding peptides based on multiple scaffolds with a wide range of affinities. The quantitative biophysical data provided by this approach will improve computational modeling efforts, elucidate a broad range of weak protein-SLiM interactions, and revolutionize our understanding of signaling networks.
View details for DOI 10.7554/eLife.40499
View details for PubMedID 31282865
Live imaging of Aiptasia larvae, a model system for coral and anemone bleaching, using a simple microfluidic device.
2019; 9 (1): 9275
Coral reefs, and their associated diverse ecosystems, are of enormous ecological importance. In recent years, coral health has been severely impacted by environmental stressors brought on by human activity and climate change, threatening the extinction of several major reef ecosystems. Reef damage is mediated by a process called 'coral bleaching' where corals, sea anemones, and other cnidarians lose their photosynthetic algal symbionts (family Symbiodiniaceae) upon stress induction, resulting in drastically decreased host energy harvest and, ultimately, coral death. The mechanism by which this critical cnidarian-algal symbiosis is lost remains poorly understood. The larvae of the sea anemone, Exaiptasia pallida (commonly referred to as 'Aiptasia') are an attractive model organism to study this process, but they are large (100 mm in length, 75 mm in diameter), deformable, and highly motile, complicating long-term imaging and limiting study of this critical endosymbiotic relationship in live organisms. Here, we report 'Traptasia', a simple microfluidic device with multiple traps designed to isolate and image individual, live larvae of Aiptasia and their algal symbionts over extended time courses. Using a trap design parameterized via fluid flow simulations and polymer bead loading tests, we trapped Aiptasia larvae containing algal symbionts and demonstrated stable imaging for >10 hours. We visualized algae within Aiptasia larvae and observed algal expulsion under an environmental stressor. To our knowledge, this device is the first to enable time-lapsed, high-throughput live imaging of cnidarian larvae and their algal symbionts and, in further implementation, could provide important insights into the cellular mechanisms of cnidarian bleaching under different environmental stressors. The 'Traptasia' device is simple to use, requires minimal external equipment and no specialized training to operate, and can easily be adapted using the trap optimization data presented here to study a variety of large, motile organisms.
View details for DOI 10.1038/s41598-019-45167-2
View details for PubMedID 31239506
- An open-source software analysis package for Microspheres with Ratiometric Barcode Lanthanide Encoding (MRBLEs) PLOS ONE 2019; 14 (3)
- A Microfluidics-Based Assay for Mapping Connectivity in Highly Proficient Enzymes Reveals Functional Modularity CELL PRESS. 2019: 66A
- A High-Throughput Platform for Probing Mechanisms of Transcription Factor-DNA Binding CELL PRESS. 2019: 502A
- Deep Learning Models Explore the Structural Effects of Transcription Factor-DNA Complexes on Binding Specificity CELL PRESS. 2019: 503A
- Bringing Enzymology into the Genomic Era: Developing and Deploying New Tools to Quantitatively Map Functional Connections Throughout an Enzyme CELL PRESS. 2019: 23A
Satb1 integrates DNA binding site geometry and torsional stress to differentially target nucleosome-dense regions.
2019; 10 (1): 3221
The Satb1 genome organizer regulates multiple cellular and developmental processes. It is not yet clear how Satb1 selects different sets of targets throughout the genome. Here we have used live-cell single molecule imaging and deep sequencing to assess determinants of Satb1 binding-site selectivity. We have found that Satb1 preferentially targets nucleosome-dense regions and can directly bind consensus motifs within nucleosomes. Some genomic regions harbor multiple, regularly spaced Satb1 binding motifs (typical separation ~1 turn of the DNA helix) characterized by highly cooperative binding. The Satb1 homeodomain is dispensable for high affinity binding but is essential for specificity. Finally, we find that Satb1-DNA interactions are mechanosensitive. Increasing negative torsional stress in DNA enhances Satb1 binding and Satb1 stabilizes base unpairing regions against melting by molecular machines. The ability of Satb1 to control diverse biological programs may reflect its ability to combinatorially use multiple site selection criteria.
View details for DOI 10.1038/s41467-019-11118-8
View details for PubMedID 31324780
micrIO: an open-source autosampler and fraction collector for automated microfluidic input-output.
Lab on a chip
Microfluidic devices are an enabling technology for many labs, facilitating a wide range of applications spanning high-throughput encapsulation, molecular separations, and long-term cell culture. In many cases, however, their utility is limited by a 'world-to-chip' barrier that makes it difficult to serially interface samples with these devices. As a result, many researchers are forced to rely on low-throughput, manual approaches for managing device input and output (IO) of samples, reagents, and effluent. Here, we present a hardware-software platform for automated microfluidic IO (micrIO). The platform, which is uniquely compatible with positive-pressure microfluidics, comprises an 'AutoSipper' for input and a 'Fraction Collector' for output. To facilitate widespread adoption, both are open-source builds constructed from components that are readily purchased online or fabricated from included design files. The software control library, written in Python, allows the platform to be integrated with existing experimental setups and to coordinate IO with other functions such as valve actuation and assay imaging. We demonstrate these capabilities by coupling both the AutoSipper and Fraction Collector to two microfluidic devices: a simple, valved inlet manifold and a microfluidic droplet generator that produces beads with distinct spectral codes. Analysis of the collected materials in each case establishes the ability of the platform to draw from and output to specific wells of multiwell plates with negligible cross-contamination between samples.
View details for DOI 10.1039/c9lc00512a
View details for PubMedID 31701110
An open-source software analysis package for Microspheres with Ratiometric Barcode Lanthanide Encoding (MRBLEs).
2019; 14 (3): e0203725
Multiplexed bioassays, in which multiple analytes of interest are probed in parallel within a single small volume, have greatly accelerated the pace of biological discovery. Bead-based multiplexed bioassays have many technical advantages, including near solution-phase kinetics, small sample volume requirements, many within-assay replicates to reduce measurement error, and, for some bead materials, the ability to synthesize analytes directly on beads via solid-phase synthesis. To allow bead-based multiplexing, analytes can be synthesized on spectrally encoded beads with a 1:1 linkage between analyte identity and embedded codes. Bead-bound analyte libraries can then be pooled and incubated with a fluorescently-labeled macromolecule of interest, allowing downstream quantification of interactions between the macromolecule and all analytes simultaneously via imaging alone. Extracting quantitative binding data from these images poses several computational image processing challenges, requiring the ability to identify all beads in each image, quantify bound fluorescent material associated with each bead, and determine their embedded spectral code to reveal analyte identities. Here, we present a novel open-source Python software package (the mrbles analysis package) that provides the necessary tools to: (1) find encoded beads in a bright-field microscopy image; (2) quantify bound fluorescent material associated with bead perimeters; (3) identify embedded ratiometric spectral codes within beads; and (4) return data aggregated by embedded code and for each individual bead. We demonstrate the utility of this package by applying it towards analyzing data generated via multiplexed measurement of calcineurin protein binding to MRBLEs (Microspheres with Ratiometric Barcode Lanthanide Encoding) containing known and mutant binding peptide motifs. We anticipate that this flexible package should be applicable to a wide variety of assays, including simple bead or droplet finding analysis, quantification of binding to non-encoded beads, and analysis of multiplexed assays that use ratiometric, spectrally encoded beads.
View details for PubMedID 30901328
Diversification of DNA binding specificities enabled SREBP transcription regulators to expand the repertoire of cellular functions that they govern in fungi.
2018; 14 (12): e1007884
The Sterol Regulatory Element Binding Proteins (SREBPs) are basic-helix-loop-helix transcription regulators that control the expression of sterol biosynthesis genes in higher eukaryotes and some fungi. Surprisingly, SREBPs do not regulate sterol biosynthesis in the ascomycete yeasts (Saccharomycotina) as this role was handed off to an unrelated transcription regulator in this clade. The SREBPs, nonetheless, expanded in fungi such as the ascomycete yeasts Candida spp., raising questions about their role and evolution in these organisms. Here we report that the fungal SREBPs diversified their DNA binding preferences concomitantly with an expansion in function. We establish that several branches of fungal SREBPs preferentially bind non-palindromic DNA sequences, in contrast to the palindromic DNA motifs recognized by most basic-helix-loop-helix proteins (including SREBPs) in higher eukaryotes. Reconstruction and biochemical characterization of the likely ancestor protein suggest that an intrinsic DNA binding promiscuity in the family was resolved by alternative mechanisms in different branches of fungal SREBPs. Furthermore, we show that two SREBPs in the human commensal yeast Candida albicans drive a transcriptional cascade that inhibits a morphological switch under anaerobic conditions. Preventing this morphological transition enhances C. albicans colonization of the mammalian intestine, the fungus' natural niche. Thus, our results illustrate how diversification in DNA binding preferences enabled the functional expansion of a family of eukaryotic transcription regulators.
View details for PubMedID 30596634
High-throughput chromatin accessibility profiling at single-cell resolution.
2018; 9 (1): 3647
Here we develop a high-throughput single-cell ATAC-seq (assay for transposition of accessible chromatin) method to measure physical access to DNA in whole cells. Our approach integrates fluorescence imaging and addressable reagent deposition across a massively parallel (5184) nano-well array, yielding a nearly 20-fold improvement in throughput (up to ~1800 cells/chip, 4-5h on-chip processing time) and library preparationcost (~81 per cell) compared to prior microfluidic implementations. We apply this method to measure regulatory variation in peripheral blood mononuclear cells (PBMCs) and show robust, de novo clustering of single cells by hematopoietic cell type.
View details for PubMedID 30194434
- Discovering epistatic feature interactions from neural network models of regulatory DNA sequences OXFORD UNIV PRESS. 2018: 629–37
Comprehensive, high-resolution binding energy landscapes reveal context dependencies of transcription factor binding
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
2018; 115 (16): E3702–E3711
Transcription factors (TFs) are primary regulators of gene expression in cells, where they bind specific genomic target sites to control transcription. Quantitative measurements of TF-DNA binding energies can improve the accuracy of predictions of TF occupancy and downstream gene expression in vivo and shed light on how transcriptional networks are rewired throughout evolution. Here, we present a sequencing-based TF binding assay and analysis pipeline (BET-seq, for Binding Energy Topography by sequencing) capable of providing quantitative estimates of binding energies for more than one million DNA sequences in parallel at high energetic resolution. Using this platform, we measured the binding energies associated with all possible combinations of 10 nucleotides flanking the known consensus DNA target interacting with two model yeast TFs, Pho4 and Cbf1. A large fraction of these flanking mutations change overall binding energies by an amount equal to or greater than consensus site mutations, suggesting that current definitions of TF binding sites may be too restrictive. By systematically comparing estimates of binding energies output by deep neural networks (NNs) and biophysical models trained on these data, we establish that dinucleotide (DN) specificities are sufficient to explain essentially all variance in observed binding behavior, with Cbf1 binding exhibiting significantly more nonadditivity than Pho4. NN-derived binding energies agree with orthogonal biochemical measurements and reveal that dynamically occupied sites in vivo are both energetically and mutationally distant from the highest affinity sites.
View details for PubMedID 29588420
An Open-Source, Programmable Pneumatic Setup for Operation and Automated Control of Single- and Multi-Layer Microfluidic Devices.
2018; 3: 117–34
Microfluidic technologies have been used across diverse disciplines (e.g. high-throughput biological measurement, fluid physics, laboratory fluid manipulation) but widespread adoption has been limited in part due to the lack of openly disseminated resources that enable non-specialist labs to make and operate their own devices. Here, we report the open-source build of a pneumatic setup capable of operating both single and multilayer (Quake-style) microfluidic devices with programmable scripting automation. This setup can operate both simple and complex devices with 48 device valve control inputs and 18 sample inputs, with modular design for easy expansion, at a fraction of the cost of similar commercial solutions. We present a detailed step-by-step guide to building the pneumatic instrumentation, as well as instructions for custom device operation using our software, Geppetto, through an easy-to-use GUI for live on-chip valve actuation and a scripting system for experiment automation. We show robust valve actuation with near real-time software feedback and demonstrate use of the setup for high-throughput biochemical measurements on-chip. This open-source setup will enable specialists and novices alike to run microfluidic devices easily in their own laboratories.
View details for PubMedID 30221210
BET-seq: Binding energy topographies revealed by microfluidics and high-throughput sequencing.
Methods in cell biology
2018; 148: 229–50
Biophysical models of transcriptional regulation rely on energetic measurements of the binding affinities between transcription factors (TFs) and target DNA binding sites. Historically, assays capable of measuring TF-DNA binding affinities have been relatively low-throughput (measuring ~103 sequences in parallel) and have required significant specialized equipment, limiting their use to a handful of laboratories. Recently, we developed an experimental assay and analysis pipeline that allows measurement of binding energies between a single TF and up to 106 DNA species in a single experiment (Binding Energy Topography by sequencing, or BET-seq) (Le et al., 2018). BET-seq employs the Mechanically Induced Trapping of Molecular Interactions (MITOMI) platform to purify DNA bound to a TF at equilibrium followed by high coverage sequencing to reveal relative differences in binding energy for each sequence. While we have previously used BET-seq to refine the binding affinity landscapes surrounding high-affinity DNA consensus target sites, we anticipate this technique will be applied in future work toward measuring a wide variety of TF-DNA landscapes. Here, we provide detailed instructions and general considerations for DNA library design, performing BET-seq assays, and analyzing the resulting data.
View details for PubMedID 30473071
Discovering epistatic feature interactions from neural network models of regulatory DNA sequences.
Bioinformatics (Oxford, England)
2018; 34 (17): i629–i637
Transcription factors bind regulatory DNA sequences in a combinatorial manner to modulate gene expression. Deep neural networks (DNNs) can learn the cis-regulatory grammars encoded in regulatory DNA sequences associated with transcription factor binding and chromatin accessibility. Several feature attribution methods have been developed for estimating the predictive importance of individual features (nucleotides or motifs) in any input DNA sequence to its associated output prediction from a DNN model. However, these methods do not reveal higher-order feature interactions encoded by the models.We present a new method called Deep Feature Interaction Maps (DFIM) to efficiently estimate interactions between all pairs of features in any input DNA sequence. DFIM accurately identifies ground truth motif interactions embedded in simulated regulatory DNA sequences. DFIM identifies synergistic interactions between GATA1 and TAL1 motifs from in vivo TF binding models. DFIM reveals epistatic interactions involving nucleotides flanking the core motif of the Cbf1 TF in yeast from in vitro TF binding models. We also apply DFIM to regulatory sequence models of in vivo chromatin accessibility to reveal interactions between regulatory genetic variants and proximal motifs of target TFs as validated by TF binding quantitative trait loci. Our approach makes significant strides in improving the interpretability of deep learning models for genomics.Code is available at: https://github.com/kundajelab/dfim.Supplementary data are available at Bioinformatics online.
View details for PubMedID 30423062
- Programmable Microfluidic Synthesis of Over One Thousand Uniquely Identifiable Spectral Codes ADVANCED OPTICAL MATERIALS 2017; 5 (3)
Multi-step Variable Height Photolithography for Valved Multilayer Microfluidic Devices.
Journal of visualized experiments : JoVE
Microfluidic systems have enabled powerful new approaches to high-throughput biochemical and biological analysis. However, there remains a barrier to entry for non-specialists who would benefit greatly from the ability to develop their own microfluidic devices to address research questions. Particularly lacking has been the open dissemination of protocols related to photolithography, a key step in the development of a replica mold for the manufacture of polydimethylsiloxane (PDMS) devices. While the fabrication of single height silicon masters has been explored extensively in literature, fabrication steps for more complicated photolithography features necessary for many interesting device functionalities (such as feature rounding to make valve structures, multi-height single-mold patterning, or high aspect ratio definition) are often not explicitly outlined. Here, we provide a complete protocol for making multilayer microfluidic devices with valves and complex multi-height geometries, tunable for any application. These fabrication procedures are presented in the context of a microfluidic hydrogel bead synthesizer and demonstrate the production of droplets containing polyethylene glycol (PEG diacrylate) and a photoinitiator that can be polymerized into solid beads. This protocol and accompanying discussion provide a foundation of design principles and fabrication methods that enables development of a wide variety of microfluidic devices. The details included here should allow non-specialists to design and fabricate novel devices, thereby bringing a host of recently developed technologies to their most exciting applications in biological laboratories.
View details for DOI 10.3791/55276
View details for PubMedID 28190039
Joker de Bruijn: Sequence Libraries to Cover All k-mers Using Joker Characters
SPRINGER-VERLAG BERLIN. 2017: 389–90
View details for PubMedID 29707702
Optimized Sequence Library Design for Efficient In Vitro Interaction Mapping.
2017; 5 (3): 230–36.e5
Sequence libraries that cover all k-mers enable universal, unbiased measurements of binding to both oligonucleotides and peptides. While the number of k-mers grows exponentially in k, space on all experimental platforms is limited. Here, we shrink k-mer library sizes by using joker characters, which represent all characters in the alphabet simultaneously. We present the JokerCAKE (joker covering all k-mers) algorithm for generating a short sequence such that each k-mer appears at least p times with at most one joker character per k-mer. By running our algorithm on a range of parameters and alphabets, we show that JokerCAKE produces near-optimal sequences. Moreover, through comparison with data from hundreds of DNA-protein binding experiments and with new experimental results for both standard and JokerCAKE libraries, we establish that accurate binding scores can be inferred for high-affinity k-mers using JokerCAKE libraries. JokerCAKE libraries allow researchers to search a significantly larger sequence space using the same number of experimental measurements and at the same cost.
View details for PubMedID 28957657
View details for PubMedCentralID PMC5661997
How duplicated transcription regulators can diversify to govern the expression of nonoverlapping sets of genes
GENES & DEVELOPMENT
2014; 28 (12): 1272-1277
The duplication of transcription regulators can elicit major regulatory network rearrangements over evolutionary timescales. However, few examples of duplications resulting in gene network expansions are understood in molecular detail. Here we show that four Candida albicans transcription regulators that arose by successive duplications have differentiated from one another by acquiring different intrinsic DNA-binding specificities, different preferences for half-site spacing, and different associations with cofactors. The combination of these three mechanisms resulted in each of the four regulators controlling a distinct set of target genes, which likely contributed to the adaption of this fungus to its human host. Our results illustrate how successive duplications and diversification of an ancestral transcription regulator can underlie major changes in an organism's regulatory circuitry.
View details for DOI 10.1101/gad.242271.114
View details for Web of Science ID 000337991000002
View details for PubMedID 24874988
Structure of the transcriptional network controlling white-opaque switching in Candida albicans
2013; 90 (1): 22-35
The human fungal pathogen Candida albicans can switch between two phenotypic cell types, termed 'white' and 'opaque'. Both cell types are heritable for many generations, and the switch between the two types occurs epigenetically, that is, without a change in the primary DNA sequence of the genome. Previous work identified six key transcriptional regulators important for white-opaque switching: Wor1, Wor2, Wor3, Czf1, Efg1, and Ahr1. In this work, we describe the structure of the transcriptional network that specifies the white and opaque cell types and governs the ability to switch between them. In particular, we use a combination of genome-wide chromatin immunoprecipitation, gene expression profiling, and microfluidics-based DNA binding experiments to determine the direct and indirect regulatory interactions that form the switch network. The six regulators are arranged together in a complex, interlocking network with many seemingly redundant and overlapping connections. We propose that the structure (or topology) of this network is responsible for the epigenetic maintenance of the white and opaque states, the switching between them, and the specialized properties of each state.
View details for DOI 10.1111/mmi.12329
View details for Web of Science ID 000324950600003
View details for PubMedID 23855748
Microfluidic affinity and ChIP-seq analyses converge on a conserved FOXP2-binding motif in chimp and human, which enables the detection of evolutionarily novel targets
NUCLEIC ACIDS RESEARCH
2013; 41 (12): 5991-6004
The transcription factor forkhead box P2 (FOXP2) is believed to be important in the evolution of human speech. A mutation in its DNA-binding domain causes severe speech impairment. Humans have acquired two coding changes relative to the conserved mammalian sequence. Despite intense interest in FOXP2, it has remained an open question whether the human protein's DNA-binding specificity and chromatin localization are conserved. Previous in vitro and ChIP-chip studies have provided conflicting consensus sequences for the FOXP2-binding site. Using MITOMI 2.0 microfluidic affinity assays, we describe the binding site of FOXP2 and its affinity profile in base-specific detail for all substitutions of the strongest binding site. We find that human and chimp FOXP2 have similar binding sites that are distinct from previously suggested consensus binding sites. Additionally, through analysis of FOXP2 ChIP-seq data from cultured neurons, we find strong overrepresentation of a motif that matches our in vitro results and identifies a set of genes with FOXP2 binding sites. The FOXP2-binding sites tend to be conserved, yet we identified 38 instances of evolutionarily novel sites in humans. Combined, these data present a comprehensive portrait of FOXP2's-binding properties and imply that although its sequence specificity has been conserved, some of its genomic binding sites are newly evolved.
View details for DOI 10.1093/nar/gkt259
View details for Web of Science ID 000321057100012
View details for PubMedID 23625967
Identification and characterization of a previously undescribed family of sequence-specific DNA-binding domains
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
2013; 110 (19): 7660-7665
Sequence-specific DNA-binding proteins are among the most important classes of gene regulatory proteins, controlling changes in transcription that underlie many aspects of biology. In this work, we identify a transcriptional regulator from the human fungal pathogen Candida albicans that binds DNA specifically but has no detectable homology with any previously described DNA- or RNA-binding protein. This protein, named White-Opaque Regulator 3 (Wor3), regulates white-opaque switching, the ability of C. albicans to switch between two heritable cell types. We demonstrate that ectopic overexpression of WOR3 results in mass conversion of white cells to opaque cells and that deletion of WOR3 affects the stability of opaque cells at physiological temperatures. Genome-wide chromatin immunoprecipitation of Wor3 and gene expression profiling of a wor3 deletion mutant strain indicate that Wor3 is highly integrated into the previously described circuit regulating white-opaque switching and that it controls a subset of the opaque transcriptional program. We show by biochemical, genetic, and microfluidic experiments that Wor3 binds directly to DNA in a sequence-specific manner, and we identify the set of cis-regulatory sequences recognized by Wor3. Bioinformatic analyses indicate that the Wor3 family arose more recently in evolutionary time than most previously described DNA-binding domains; it is restricted to a small number of fungi that include the major fungal pathogens of humans. These observations show that new families of sequence-specific DNA-binding proteins may be restricted to small clades and suggest that current annotations--which rely on deep conservation--underestimate the fraction of genes coding for transcriptional regulators.
View details for DOI 10.1073/pnas.1221734110
View details for Web of Science ID 000319327700041
View details for PubMedID 23610392
Programmable microfluidic synthesis of spectrally encoded microspheres
LAB ON A CHIP
2012; 12 (22): 4716-4723
Spectrally encoded fluorescent beads are an attractive platform for assay miniaturization and multiplexing in the biological sciences. Here, we synthesize hydrophilic PEG-acrylate polymer beads encoded with lanthanide nanophosphors using a fully automated microfluidic synthesis device. These beads are encoded by including varying amounts of two lanthanide nanophosphors relative to a third reference nanophosphor to generate 24 distinct ratios. These codes differ by less than 3% from their target values and can be distinguished from each other with an error rate of <0.1%. The encoded bead synthesis strategy we have used is readily extensible to larger numbers of codes, potentially up to millions, providing a new platform technology for assay multiplexing.
View details for DOI 10.1039/c2lc40699c
View details for Web of Science ID 000310865200017
View details for PubMedID 23042484
Basic leucine zipper transcription factor Hac1 binds DNA in two distinct modes as revealed by microfluidic analyses
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
2012; 109 (45): E3084-E3093
A quantitative understanding of how transcription factors interact with genomic target sites is crucial for reconstructing transcriptional networks in vivo. Here, we use Hac1, a well-characterized basic leucine zipper (bZIP) transcription factor involved in the unfolded protein response (UPR) as a model to investigate interactions between bZIP transcription factors and their target sites. During the UPR, the accumulation of unfolded proteins leads to unconventional splicing and subsequent translation of HAC1 mRNA, followed by transcription of UPR target genes. Initial candidate-based approaches identified a canonical cis-acting unfolded protein response element (UPRE-1) within target gene promoters; however, subsequent studies identified a large set of Hac1 target genes lacking this UPRE-1 and containing a different motif (UPRE-2). Using a combination of unbiased and directed microfluidic DNA binding assays, we established that Hac1 binds in two distinct modes: (i) to short (6-7 bp) UPRE-2-like motifs and (ii) to significantly longer (11-13 bp) extended UPRE-1-like motifs. Using a genetic screen, we demonstrate that a region of extended homology N-terminal to the basic DNA binding domain is required for this dual site recognition. These results establish Hac1 as the first bZIP transcription factor known to adopt more than one binding mode and unify previously conflicting and discrepant observations of Hac1 function into a cohesive model of UPR target gene activation. Our results also suggest that even structurally simple transcription factors can recognize multiple divergent target sites of very different lengths, potentially enriching their downstream target repertoire.
View details for DOI 10.1073/pnas.1212457109
View details for Web of Science ID 000311156700009
View details for PubMedID 23054834
View details for PubMedCentralID PMC3494901
Systematic characterization of feature dimensions and closing pressures for microfluidic valves produced via photoresist reflow
LAB ON A CHIP
2012; 12 (21): 4287-4295
Multilayer soft lithography (MSL) provides a convenient and low-cost method for fabricating poly(dimethyl siloxane) (PDMS) microfluidic devices with on-chip valves for automated and precise control of fluid flow. MSL casting molds for flow channels typically incorporate small patches of rounded positive photoresist at valve locations to achieve the rounded cross-sectional profile required for these valves to function properly. Despite the importance of these rounded features for device performance, a comprehensive characterization of how the rounding process affects feature dimensions and closing pressures has been lacking. Here, we measure valve dimensions both before and after rounding and closing pressures for 120 different valve widths and lengths at post-rounding heights between 15 and 84 μm, for a total of 1200 different geometries spanning a wide range of useful sizes. We find that valve height and width after rounding depend strongly on valve aspect ratios, with these effects becoming more pronounced for taller and narrower features. Based on the measured data, we provide a simple fitted model and an online tool for estimating the pre-rounding dimensions needed to achieve desired post-rounding dimensions. We also find that valve closing pressures are well explained by modelling valve membranes in a manner analogous to a suspension bridge, shedding new light on device physics and providing a practical model for estimating closing pressures during device design.
View details for DOI 10.1039/c2lc40414a
View details for Web of Science ID 000310916100012
View details for PubMedID 22930180
De novo identification and biophysical characterization of transcription-factor binding sites with microfluidic affinity analysis
2010; 28 (9): 970-976
Gene expression is regulated in part by protein transcription factors that bind target regulatory DNA sequences. Predicting DNA binding sites and affinities from transcription factor sequence or structure is difficult; therefore, experimental data are required to link transcription factors to target sequences. We present a microfluidics-based approach for de novo discovery and quantitative biophysical characterization of DNA target sequences. We validated our technique by measuring sequence preferences for 28 Saccharomyces cerevisiae transcription factors with a variety of DNA-binding domains, including several that have proven difficult to study by other techniques. For each transcription factor, we measured relative binding affinities to oligonucleotides covering all possible 8-bp DNA sequences to create a comprehensive map of sequence preferences; for four transcription factors, we also determined absolute affinities. We expect that these data and future use of this technique will provide information essential for understanding transcription factor specificity, improving identification of regulatory sites and reconstructing regulatory interactions.
View details for DOI 10.1038/nbt.1675
View details for Web of Science ID 000281719100024
View details for PubMedID 20802496
View details for PubMedCentralID PMC2937095
Individual dimers of the mitotic kinesin motor Eg5 step processively and support substantial loads in vitro
NATURE CELL BIOLOGY
2006; 8 (5): 470-U89
Eg5, a member of the kinesin superfamily of microtubule-based motors, is essential for bipolar spindle assembly and maintenance during mitosis, yet little is known about the mechanisms by which it accomplishes these tasks. Here, we used an automated optical trapping apparatus in conjunction with a novel motility assay that employed chemically modified surfaces to probe the mechanochemistry of Eg5. Individual dimers, formed by a recombinant human construct Eg5-513-5His, stepped processively along microtubules in 8-nm increments, with short run lengths averaging approximately eight steps. By varying the applied load (with a force clamp) and the ATP concentration, we found that the velocity of Eg5 was slower and less sensitive to external load than that of conventional kinesin, possibly reflecting the distinct demands of spindle assembly as compared with vesicle transport. The Eg5-513-5His velocity data were described by a minimal, three-state model where a force-dependent transition follows nucleotide binding.
View details for DOI 10.1038/ncb1394
View details for Web of Science ID 000237299400010
View details for PubMedID 16604065
View details for PubMedCentralID PMC1523314
Eg5 steps it up!
Understanding how molecular motors generate force and move microtubules in mitosis is essential to understanding the physical mechanism of cell division. Recent measurements have shown that one mitotic kinesin superfamily member, Eg5, is mechanically processive and capable of crosslinking and sliding microtubules in vitro. In this review, we highlight recent work that explores how Eg5 functions under load, with an emphasis on the nanomechanical properties of single enzymes.
View details for DOI 10.1186/1747-1028-1-31
View details for Web of Science ID 000207723600031
View details for PubMedID 17173688
View details for PubMedCentralID PMC1716758
Simultaneous, coincident optical trapping and single-molecule fluorescence
2004; 1 (2): 133-139
We constructed a microscope-based instrument capable of simultaneous, spatially coincident optical trapping and single-molecule fluorescence. The capabilities of this apparatus were demonstrated by studying the force-induced strand separation of a dye-labeled, 15-base-pair region of double-stranded DNA (dsDNA), with force applied either parallel ('unzipping' mode) or perpendicular ('shearing' mode) to the long axis of the region. Mechanical transitions corresponding to DNA hybrid rupture occurred simultaneously with discontinuous changes in the fluorescence emission. The rupture force was strongly dependent on the direction of applied force, indicating the existence of distinct unbinding pathways for the two force-loading modes. From the rupture force histograms, we determined the distance to the thermodynamic transition state and the thermal off rates in the absence of load for both processes.
View details for DOI 10.1038/NMETH714
View details for Web of Science ID 000226753800017
View details for PubMedID 15782176
View details for PubMedCentralID PMC1483847
Stepping and stretching - How kinesin uses internal strain to walk processively
JOURNAL OF BIOLOGICAL CHEMISTRY
2003; 278 (20): 18550-18556
The ability of kinesin to travel long distances on its microtubule track without dissociating has led to a variety of models to explain how this remarkable degree of processivity is maintained. All of these require that the two motor domains remain enzymatically "out of phase," a behavior that would ensure that, at any given time, one motor is strongly attached to the microtubule. The maintenance of this coordination over many mechanochemical cycles has never been explained, because key steps in the cycle could not be directly observed. We have addressed this issue by applying several novel spectroscopic approaches to monitor motor dissociation, phosphate release, and nucleotide binding during processive movement by a dimeric kinesin construct. Our data argue that the major effect of the internal strain generated when both motor domains of kinesin bind the microtubule is to block ATP from binding to the leading motor. This effect guarantees the two motor domains remain out of phase for many mechanochemical cycles and provides an efficient and adaptable mechanism for the maintenance of processive movement.
View details for DOI 10.1074/jbc.M300849200
View details for Web of Science ID 000182838300126
View details for PubMedID 12626516
View details for PubMedCentralID PMC1533991
Combined optical trapping and single-molecule fluorescence.
Journal of biology
2003; 2 (1): 6-?
Two of the mainstay techniques in single-molecule research are optical trapping and single-molecule fluorescence. Previous attempts to combine these techniques in a single experiment - and on a single macromolecule of interest - have met with little success, because the light intensity within an optical trap is more than ten orders of magnitude greater than the light emitted by a single fluorophore. Instead, the two techniques have been employed sequentially, or spatially separated by distances of several micrometers within the sample, imposing experimental restrictions that limit the utility of the combined method. Here, we report the development of an instrument capable of true, simultaneous, spatially coincident optical trapping and single-molecule fluorescence.We demonstrate the capability of the apparatus by studying force-induced strand separation of a rhodamine-labeled, 15 base-pair segment of double-stranded DNA, with force applied perpendicular to the axis of the DNA molecule. As expected, we observed abrupt mechanical transitions corresponding to the unzipping of DNA at a critical force. Transitions occurred concomitant with changes in the fluorescence of dyes attached at the duplex ends, which became unquenched upon strand separation.Through careful optical design, the use of high-performance spectral notch filters, a judicious choice of fluorophores, and the rapid acquisition of data gained by computer-automating the experiment, it is possible to perform combined optical trapping and single-molecule fluorescence. This opens the door to many types of experiment that employ optical traps to supply controlled external loads while fluorescent molecules report concurrent information about macromolecular structure.
View details for PubMedID 12733997