Slide 1:microRNA PROFILING USING A HIGH PERFORMANCE, FLEXIBLE μPARAFLO®  BIOCHIP PLATFORM Christoph Eicken, PhD Head of Technical Services - Microarrays


Slide 2:Agenda miRNAIntro Q & A


Slide 3:LC Sciences headquarters in Houston, Texas - USA offices in USA & China, representatives inEurope, Japan, Korea, & India customers in >40 countries first miRNA array service provider (since 2005)


Slide 4:A Short History of microRNA Discovery 1993 - Lin-4 was shown to encode two small RNA molecules (not protein) and control developmental timing in C. elegans through negative regulation of lin-14 gene. (Lee et al.) Challenge to the Central Dogma of Biology DNA > transcription > RNA > translation > protein 2000 - Let-7 (another non-coding gene) was shown to regulate development in worms and a homolog of let-7 was identified in humans and Drosophila. (Reinhart et al., Slack et al.) 2004 – Neuroscience A novel microRNA mediates left−right asymmetry in a pair of C. elegans chemosensory neurons. (Chang et al.) 2005 – Three studies published that definitively link microRNAs with cancer Classification of poorly differentiated tumours using miRNA expression profiles. (Lu et al.) microRNAs can modulate tumor formation. (He et al.) c-Myc (whose dysregulated expression is common abnormalities in human malignancy) activates expression of a cluster of six miRNAs on human chromosome. (O’Donnel et al.) 2006 – microRNA linked with Heart Disease miRNA expression profiling studies demonstrate that expression levels of specific miRNAs change in diseased human hearts, pointing to their involvement in cardiomyopathies (van Rooij et al.)


Slide 5:Summary Of What We Know 17-25nt short RNA – Precursors consist of predictable stem and loop structure - highly conserved Bind to complementary mRNA molecules and act as negative regulators of translation High Copy number Expression is tissue (and developmental stage) specific Currently - 10581 mature miRNAs across ~115 plant and animal species and >12 viruses that infect them (miRBase 14.0, September 2009) Sequence Database miRBase - updated regularly as new sequences are experimentally verified: http://mirbase.org/ Mechanism is far reaching and complex – Each miRNA may control many genes and it is estimated that miRNAs regulate expression of up to 1/3 of all human genes. Operate by one of two hypothesized mechanisms: – Complete pairing mRNA is degraded Predominant in plants – Imperfect pairing translation is repressed but mRNA remains intact Predominant in animals


Slide 6:microRNA Processing pri-miRNA = primary microRNA transcriptpre-miRNA = precursor microRNAmiRNA* = antisense microRNAmiRISC = microRNA-induced silencing complex


Slide 7:microRNA Biogenesis Pathway (A) Animal and (B) Plant miRNA biogenesis. Mature miRNAs are indicated in red and miRNA* strands are in blue. Du, T. et al. Development 2005;132:4645-4652


Slide 8:The structure of human pri-miRNAs miRNAs with their own transcription units miR-21 the polycistronic miR-17–92-1 cluster (Cai et al., 2004; He et al., 2005) Du, T. et al. Development 2005;132:4645-4652 ~ 1/3 Reside inside introns ~ 2/3 independent transcription units Often found in clusters Many times near the genes they regulate or inside them


Slide 9:The structure of human pri-miRNAs miRNAs that are transcribed with other genes miR-15a 16-1 resides in the intron of a non-coding RNA (ncRNA) (Calin et al., 2004) miR-106b 93 25 lies in the intron of a protein-coding RNA (Rodriguez et al., 2004) miR-155 is found in the exon of a ncRNA (Eis et al., 2005), whereas miR-198 is in the exon of a protein-coding mRNA (Cullen, 2004) Du, T. et al. Development 2005;132:4645-4652


Slide 10:microRNA hits the Headlines


Slide 11:microRNA Related Publications Source: PubMed


Slide 12:Microfluidic Array Platform μParaflo®  Microfluidics Chip 10 µl total volume 4000 features


Slide 13:The Core: 3 Levels of Flexibility Chemistry Digital Photolithography Microfluidics


Slide 14:Chemistry Does not require electrochemical surface Nor specialty monomers Standard DMT chemistry allows use ofwide array of (non-regular) building blocks The Core: 3 Levels of Flexibility


Slide 15:3 Levels of Flexibility Digital Photolithography custom sequences on demand synthesis


Slide 16:3 Levels of Flexibility Microfluidics multiple washing steps T-variation using the same chip, keeping all other parameters constant


Slide 17:Si Substrate Cover Glass Distribution Channel Fluid Channel Reaction Chamber Light Beam Photogenerated Reagent μParaflo®  Microfluidics Chip


Slide 18:DNA DNA array OligoMix® Modified Nucleotides miRNA array ncRNA array Peptides Kinase array Protein binding array Custom arrays Probe Diversity


Slide 19:miRBase Version # of sequences (all species) miRBase Synchronicity


Slide 20:Probe Content: miRBase an Beyond miRBase 14.0


Slide 21:Microfluidics vs. Spotted Arrays - Raw Data


Slide 22:Microfluidics vs. Spotted Arrays - Raw Data


Slide 23:Spot Uniformity Enhanced Mass Transfer: improving sensitivity High Reproducibility Closed System:no dye “bleaching” Incremental Stringency Wash: higher specificity Stable for years Microfluidics vs. Spotted Arrays


Slide 24:Specificity: Optimized RNA Hybridization Probes Sample: 20 mer RNA control spiked into total RNA sample Signal Intensity


Slide 25:Brain qRT-PCR Validation Dual sample array using mouse brain and mouse thymus RNA samples from Ambion.  Results are compared with QPCR data for the same two RNA samples, also purchased from Ambion, published by Applied Biosystems (ABI)*. The comparison data includes all 12 microRNA transcripts published by ABI. * Data is obtained from http://www.appliedbiosystems.com


Slide 26:To Sum It Up: Microfluidic Microarray Platform:Reliable and accurate results as validated with qPCR and Northern Blot Optimized RNA Hybridization Probes:Designed with optimized Tms to ensure uniform hybridization affinity under high stringency hybridization conditions enhancing both, sensitivity and specificity of the probes Most Current Content:Not 90% or 95% coverage but always 100% of experimentally verified miRNA sequences on all our arrays Complete Content Flexibility:The choice is yours! No Limitation on Use of Data:Full access and control over the results generated from your experiment


Slide 27:Customer Publications: 75 100+ Cancer Research Cardiovascular Research Neuroscience Small RNA Discovery Reproductive Biology Plant Science Virology (HIV, Epstein-Barr) Stem Cell Research Endocrinology


Slide 28:microRNA Related Publications Source: PubMed LC Science Customers PubMed


Slide 29:Pseudo color images Biomarker Discovery / Profiling: Differential Expression - One Dual Sample Chip control (Cy3) vs treated (Cy5) Cy3/Cy5 ratio


Slide 30:Sample Data Repeats


Slide 31:Raw Data in Publications


Slide 32:Control / Treated Biological repeats t-Test Multi-array normalization and clustering analysis Array assay Differentiated miRNAs of Biological & Statistical Significance - Multiple Chips


Slide 33:Disease - Marker identification(Olga, Olson, Cameron) Cameron JE, Yin Q, Fewell C, Lacey M, McBride J, Wang X, Lin Z, Schaefer BC, Flemington EK. (2008)The Epstein-Barr Virus latent membrane protein 1 (LMP1) induces cellular microRNA-146a, a modulator of lymphocyte signaling pathways. J Virol [Epub ahead of print]. Disease Marker Discovery


Slide 34:Cameron JE, Yin Q, Fewell C, Lacey M, McBride J, Wang X, Lin Z, Schaefer BC, Flemington EK. (2008)The Epstein-Barr Virus latent membrane protein 1 (LMP1) induces cellular microRNA-146a, a modulator of lymphocyte signaling pathways. J Virol [Epub ahead of print]. Experiment included 5 biological repeats/group Probes for human (hsa) & virus (kshv) miRNA were used on the same chip


Slide 35:From Biomarkers to Therapeutics Tavazoie SF, Alarcón C, Oskarsson T, Padua D, Wang Q, Bos PD, Gerald WL, Massagué J. (2008) Endogenous human microRNAs that suppress breast cancer metastasis. Nature 451(7175), 147-52.


Slide 36:Discovery Applications: Potential miRNA or other small regulatory RNA genes array result


Slide 37:Discovery Applications: mRNA 3’ 3’ UTR Protein miRNA Target Screening


Slide 38:Vagin VV, Sigova A, Li C, Seitz H, Gvozdev V, Zamore PD. (2006) A distinct small RNA pathway silences selfish genetic elements in the germline. Science 313(5785), 320-324. Cummins JM, He Y, Leary RJ, Pagliarini R, Diaz L.A Jr, Sjoblom T, Barad O, Bentwich Z, Szafranska AE, Labourier E, et al. (2006) The colorectal microRNAome. Proc Natl Acad Sci USA 103, 3687-3692. Discovery Applications:


Slide 39:Workflow


Slide 40:microRNA Microarray Service - Reports


Slide 41:Which Probe Content Should I Chose? Most current miRBase content If correlation to older data needed:chose same probe content single vs multi species chips:less can be more... custom probes:(predicted sequences, controls etc.) proprietary probes (additional paperwork, who will benefit from those data?)


Slide 42:For differential studies, such as normal versus disease, untreated versus treated, and time course experiments, it is very important to have biological repeats in the experiment. Statistic identification of biologically significant calls can only be made based on comparing miRNA expression difference between different sample groups with the expression variations among samples within corresponding groups. Statistic significance is determined by a statistic test, such as T-test. The mathematical formulation of the T-test is T function where, A and B are the average value of group A and B, respectively; SA and SB are the standard deviations of group A and B, respectively; and nA and nB are the number of samples in group A and B, respectively. The higher the T function, the higher the probability of groups A and B being statistically different. Obviously, a statistic test can be performed only when each group contains more than one sample. Importance of biological repeats _ _


Slide 43:Number of biological repeats needed The higher the biological variations are expected the more biological repeats should be used. Cell line samples: preferably 3 or more, at least 2 Lab animals: preferably 4 or more, at least 3 Human samples: 10 or more


Slide 44:Technical repeats Generally there is no need for technical repeats. It is much more technically effective and financially efficient to use biological repeats instead of technical repeats. In most cases, our assay variations are less than biological-repeat variations. Pooled biological samples Generally, pooled samples are not recommended. When samples are pooled, critical formation on sample-to-sample variations (SA and SB in equation (1)) within the same groups are lost and therefore identification of biologically significant miRNA differentials may no longer be possible.


Slide 45:Dual sample vs single sample assay A dual-sample assay is used when two samples are compared with each other. The assay does not involve chip-to-chip variations and therefore may reveal very small differences between the two samples. The assay is especially suitable when paired samples are studied, such as diseased tissue being compared with an adjacent non-diseased tissue. The assay also has an advantage of lower per-sample cost than single sample assay does. However, design considerations should be given to minimize dye-related bias. A single-sample assay is used when multiple independent samples are compared with each other. The assay has the advantage of free of dye-related bias although it has a higher per-sample cost.


Slide 46:Minimize dye bias in dual-sample assays Dye swap can be used to eliminate false calls due to dye-related bias in dual sample assays. A good design should have half members of each group labeled with Cy3 and the other half labeled with Cy5. Example 1: 4 samples (A1, A2, B1, and B2) from an experiment of comparing untreated versus treated samples with 2 biological repeats for each sample group.


Slide 47:Minimize dye bias in dual-sample assays Example 2: 12 samples from an experiment involving three sample groups of untreated (A1-A4), treated at dosage 1 (B1-B4), and treated at dosage 2 (C1-C4).


Slide 48:Normalizaton LOWESS (Locally weighted scatterplot smoothing)is one of many "modern" modeling methods that build on "classical" methods, such as linear and nonlinear least squares regression. LOWESS combines much of the simplicity of linear least squares regression with the flexibility of nonlinear regression. It does this by fitting simple models to localized subsets of the data to build up a function that describes the deterministic part of the variation in the data, point by point. In fact, one of the chief attractions of this method is that the data analyst is not required to specify a global function of any form to fit a model to the data, only to fit segments of the data.The trade-off for these features is increased computation. Your controls (optional) Our controls (optional)


Slide 49:microRNA Microarray Service Featuring µParaflo™ Biochip Technology 100% CURRENT miRBASE PROBE CONTENT NUMEROUS CUSTOMER PUBLICATIONS OPTIMIZED RNA HYBRIDIZATION PROBES COST EFFECTIVE 1-STOP SOLUTION COMPLETE CONTENT FLEXIBILITY