microRNA Microarray Service

Our Comprehensive Service

At LC Sciences, we offer a genome-wide microRNA expression profiling service utilizing a microarray detection system that was developed specifically for microRNA detection.  Send us your total RNA sample and we’ll perform all the necessary functions from sample QC through data analysis.  We can perform microarray analysis on a single sample to create a simple expression profile or we can hybridize two samples to the same microarray for “dual sample” analysis.  This is very useful whenever comparison of two samples is needed such as wildtype vs. mutant or samples treated in two different ways.

Our “Total RNA to Data” comprehensive service includes: Sample QC, Sample Preparation and Labeling, microRNA Detection (hybridization to a µParaflo^® microfluidics chip), Array Scan and Data Extraction, and Full Data Analysis.

In about 2-3 weeks you will receive: original and processed scan images, array layout and a list of sequences,  raw and processed data, a list of differentially expressed transcripts (dual sample arrays only), and a summary of the results.

Probe Content

Our microRNA arrays cover all species for which sequence data are available in the miRBase Sequence Database and the Plant MicroRNA Database.  Although these sequence databases are being continually updated as new sequences are experimentally verified, the contents of our standard arrays are updated in synchronization with the databases.  This synchronization is made possible by our flexible µParaflo^® microfluidic chip technology.  We continuously offer the flexibility of sequence selection.  Customers who want to combine miRNAs of different species and/or add custom sequences to standard arrays may do so by requesting custom arrays.  When the number of added sequences is below 100, the addition will be free of additional charge.

Special Announcement – miRBase Version 19 – Now Available!

August 6th , 2012 – The miRBase sequence database was just updated to version 19. Release 19 of the database contains 21,264 entries representing hairpin precursor miRNAs, expressing 25,141 mature miRNA products, in 193 species. As compared to miRBase version 18 a total of 3,171new hairpin sequences and 3,625novel mature products have been added.

Custom microRNA Microarrays

These are not off-the-shelf spotted arrays!  The flexible µParaflo^® microfluidic chip technology enables us to produce custom synthesized microarrays when ordered. (vs. an off-the-shelf spotted array) Therefore, you can add any sequence of your design to our standard microRNA probe content.  The contents of our standard microarrays are customizable by adding up to 100 customer specified sequences at no cost.  Customization of your array will not cause any delay in data delivery as we synthesize all microarrays to order.  Simply enter your custom sequences on the Sample Submission Form.

Customizable features include – sequence design, varying chain lengths, chip layout, synthesis chemistry, and more! Each µParaflo^® microfluidic chip has room for thousands of sequences of your design.
Add sequences for various applications:

• Screen for new microRNAs by adding predicted mature microRNA sequences or perform sequence tiling along certain sequences sections.
• Combine microRNA sequences of different species to identify cross-species conservations.
• Add controls of customer’s choice for the detection of customer-added spiking RNA sequences and use as customer-selected internal controls.
• Add probes for the detection of siRNAs and/or other small non-coding RNAs.

Data Analysis / Results

We can generally have data back to the customer about 2-3 weeks from the date we receive their total RNA sample.  Full data analysis is included with our array service so that the customer can immediately use the information derived from the experiments without any further analysis.  For each array, the customer receives:

• The original and processed microarray scan images.
• An array layout file.
• A raw intensity data file in Excel.
• A fully processed data file in Excel.
• A list of up and down regulated transcripts that are called based on a statistical analysis.
• Additionally, for each batch of samples, the customer receives a Data Summary containing a catalog of data files, images of representative regions of corresponding arrays, and descriptions of specific features of the arrays.

The result of the data analysis helps our customers to save significant down-stream cost by quickly zooming in on relevant target microRNA transcripts for further studies.  The Data Summary will be emailed and a complete data set will be burned to a CD and mailed to you.

Background Subtraction and Normalization

We have in house software for routine array data processing that follows the common practices of DNA array data treatment.^1-4 In our process, data will be corrected by subtracting background and normalized to the statistical mean of all detectable transcripts.  The data are processed in a MS Excel spreadsheet using a program routine that performs raw signal background subtraction using a local regression method (Xiaochuan Zhou, unpublished results; note that the photolithographically fabricated arrays do not have peripheral areas for background values)   Data normalization, using a cyclic LOWESS (Locally-weighted Regression) method⁵ is used to remove system related variations, such as sample amount variations, dye labeling bias, and signal gain differences between scanners, so that biological relevant variations can be faithfully revealed.

Signal Detection and Analysis

Detected signals greater than background plus 3 times the standard deviation will be derived for each color channel; the mean and the co-variance (CV = stdev x100/replicate mean) of each probe having a detected signal will be calculated.  For two color experiments, the ratio (log transformed) of the two sets of detected signals, and p-values of the t-test, will be calculated.

Differentially detected signals are generally accepted as true when the ratios of the p value is less than 0.01.  For clustering analysis of multiple datasets, data adjustment includes data filtering, Log2 transformation, and gene centering and normalization.  Data filtering will remove clustering values from the data set (detected signals or detected ratios that are below a threshold value). Data centering and normalization will transform Log2 values using the mean and the standard deviation for individual miRNA across all samples.

Clustering

We now offer in depth clustering analysis to illustrate relationships among the data from complex microarray experiments.  We will perform clustering with a hierarchical method using average linkage and Euclidean distance metric.  The clustering data can be visualized using one of the several microarray programs, such as TIGR MeV (Multiple Experimental Viewer) (the Institute for Genomic Research).

Example Data

View and download a complete set of microRNA Microarray experiment data

Assay Performance

Through the use of the µParaflo^® microfluidic chip technology and design of probes containing proprietary chemical modifications, we have optimized this product to offer exceedingly high levels of sensitivity and specificity.

Sensitivity

Spot density is accurately controlled during production and has been optimized for maximum signal with minimal background noise. The low system noise means reliable calls for the expression differentials.  Our microRNA detection dynamic range is no less than 3.5 logs and the lower detection limit is less than 10 attomole.  These numbers are derived by using an experimental design method called the Latin Square Test.

Specificity

Very high detection specificity is ensured on every assay performed using ourµParaflo^® microfluidic chip technology.  Our probes are designed with a proprietary chemical modification that achieves enhanced binding to the short microRNAs.  These modifications also enhance specificity whereas other types of DNA modifications just offer improved binding affinity.  In this case the probes may be too “sticky” so that non-specific binding can occur.  On each chip we have multiple perfect match and mismatch QC (quality control) probes detecting spiked-in (20 mer) RNA controls which are added into every sample and co-labeled and co-hybridized with the sample to assess specificity.

Through the use of the µParaflo^® microfluidic chip technology and design of probes containing proprietary chemical modifications, we can achieve very uniform hybridization that is extremely reproducible.  Much more reproducible than is possible with a spotted array.

Performance Data

View a complete set of microRNA Microarray performance data

Microarray Platform

There are two main features that set LC Sciences apart from other microarray platforms and enable us to achieve such high quality and product reliability:  our microarrays are in situ synthesized right on a microchip using our µParaflo^® microfluidic chip technology,and our probes are designed with unique proprietary chemical modifications for enhanced sensitivity and specificity.

In situ Synthesis

A proprietary µParaflo^® microfluidic chip is used.  The microarray chip consists of thousands of three-dimensional chambers and is a closed system so dye oxidation and deterioration are not an issue!  The microfluidic technology produces a uniform distribution of the sample solutions on the array and enhances binding reactions and stringency wash processes.   In situoligonucleotide synthesis using PGA (photogenerated acid) coupled with conventional DMT chemistry means high probe quality, tight process control, and complete content flexibility.  Our advanced manufacturing process ensures highly uniform spots and high reproducibility across lots of arrays and yet permits total customization of contents on each individual array.  In comparison, spotted microarrays tend to suffer from poor spot uniformity and large spot to spot and array to array variations, which lead to large data deviations.  The spotting process requires significant up-front investment for oligo libraries and spotting equipment and permits no flexibility for content update or customization.

Probe Design

Each of our detection probes contains a coding segment and a long spacer.  The coding segment is a nucleotide sequence involving proprietary chemical modification for enhancing the sensitivity and specificity for the detection of target transcripts.  The spacer is a non-nucleotide molecule that extends the detection probe away from the substrate and therefore reduces surface effects and further enhances the binding between the probe and the target.  Probe repeats are used on each array to allow statistical analysis of the data.

The Tms of our detection probes are balanced by incorporation of chemically modified nucleotides with increased binding affinities.  These are not standard modified nucleotides that often have an undesirable “stickiness” characteristic.  We have improved detectability and specificity in our arrays compared to those made from regular DNA probes.  By varying the number of modified nucleotides in each probe, we can adjust the Tm of that probe.

Quality Control

Multiple QC steps are implemented at various stages of array manufacturing and assay processes.  Before being released for customer sample assays, each array must pass a stringent QC test involving hybridization with a group of control oligos.  Based on the reading from 16 sets of control probes spatially distributed across the array, signal intensities, spot uniformity, cross-array spot-to-spot uniformity, and perfect-match vs. mismatch specificities are thoroughly evaluated.  For the QC of the entire assay process, a fixed amount of several 20-mer RNA oligos is spiked into each customer sample as external controls.  Multiple sets of control probes are designed to detect the spiked-in controls.

1. Ball, C. A.; Sherlock, G.; Parkinson, H.; Rocca-Sera, P.; Brooksbank, C.; Causton, H. C.; Cavalieri, D.; Gaasterland, T.; Hingamp, P.; Holstege, F.; Ringwald, M.; Spellman, P.; Stoeckert, C. J., Jr.; Stewart,
   J. E.; Taylor, R.; Brazma, A.; Quackenbush, J., Standards for microarray data. Science 2002, 298, 539.
   [abstract]
2. Quackenbush, J., Computational analysis of microarray data. Nat Rev Genet 2001, 2, 418-27.
   [abstract]
3. Quackenbush, J., Microarray data normalization and transformation. Nat Genet 2002, 32 Suppl, 496-501.
   [abstract]
4. Sturn, A.; Quackenbush, J.; Trajanoski, Z., Genesis: cluster analysis of microarray data. Bioinformatics 2002, 18, 207-8.
   [abstract]
5. Bolstad, B. M.; Irizarry, R. A.; Astrand, M.; Speed, T. P., A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 2003, 19, 185-93.
   [abstract]