White paper draft 5/31/2017 The Four Most-Commonly Encountered Problems in Multiplex Panel Design Abstract The increasing abundance of genomic data facilitates the expansion of molecular diagnostic tools. Multiplex panels that simultaneously measure multiple analytes in a single run are becoming commonplace, and will only continue to increase in breadth and depth as technology advances. In this white paper we will discuss the four most frequently encountered problems in multiplex panel design: development time, development cost, expertise limitations and the use of sub-optimal tools. Introduction Developing multiplex panels is a complex endeavor; primers are constrained by the target, interactions with other primers and assay parameters. As a result, primer design is the main reason for panel failures, especially false negatives. When multiple reactions are required to work simultaneously, the complexity of primer design increases exponentially. For example, 10 primer pairs for each of 20 analytes in a panel equate to 200 forward primers and 200 reverse primers or 1020 possible multiplex solutions. Insert table to show complexity. “When there are multiple reactions working simultaneously the complexity increases exponentially. It gets to a point where it is impossible for a human being to rationally figure it out in any logical or realistic way. You can get lucky, or use trial and error but you are just fooling yourself.” Arjang Hassibi, Ph.D., CEO Insilixia. PanelPlex™ resolves the complexity of multiplex panel primer design. The cloud-based application provides consensus designs, and is easy to use for both single and large multiplex assays. PanelPlex represents the best of what DNA Software is known for: difficult targets, high-level multiplex, assay optimization and the elimination of trial-and-error approaches to help companies more efficiently build better diagnostics. Multiplex Panel Design Challenges The four most frequently encountered problems in multiplex panel design are development time, development cost, expertise limitations and the use of sub-optimal tools. Development Time…Impacts Time to Market Primer design is a critical early-stage step in multiplex panel development. Depending on the complexity of the panel, trial-and-error approaches can take up to six months, and involve multiple employees working on not only the in-silico design but also the wet laboratory empirical testing. If the initial designs fail, as they often do, the trial-and-error cycle repeats itself repetitively until an acceptable end solution is determined. 1 White paper draft 5/31/2017 Even after the assay is introduced, the trial-and-error cycle may begin again if initial broader market usage uncovers false negatives that mandate test revision. One approach to primer design in infectious diseases is to pull and align genomes, then to hand-pick homologous locations across the different subtypes. Although the approach can work it is very time consuming and requires computational power that is not always accessible. Computer centers may not be available or may be company-wide shared resources with limited availability. As many sequences as possible are identified and the differences assessed to locate the most conserved region in a part of genome that is highly important to how the organism functions. Unfortunately, these regions are also often targeted by drug companies, and should be avoided to extend the assay lifespan. Drugs make organisms that cause infectious diseases to mutate faster, and may influence assay revisions cycles. “A lot of time and effort goes into designing assays yet a lot are on version 2 or 3. Sometimes after a product is introduced you start seeing false negatives and portions of the assay require rework.” Jaime Prout, Developmental Scientist II, Beckman Coulter If a multiplex assay design is critical to a company’s mission, more efficient approaches are needed to accelerate discovery work. Development Cost…Computational versus Empirical Time and effort for trial-and-error approaches have a significant cost. Not only are there fully-burdened employee costs, but also associated laboratory and reagent costs in addition to lost revenue due to market-entry delay (Table X). In a highly-competitive market, product introduction delays can mean the difference between securing market leadership as opposed to the number two, three or lower spots. “Relying more on computer-aided designs instead of empirical work drastically reduces labor and reagent costs. We have proven that we can have 70% of the job done with the computational design and the remaining 30% is empirical effort. That definitely reduces our costs.” Arjang Hassibi, Ph.D., CEO Insilixia. Table X Development Cost 3 – 5 Fully-burdened employees Access to computational power Laboratory space, reagents and supplies Lost revenues due to delayed market entry Crowd computing eliminates time-consuming trial-and-error research, cut costs and increases productivity. This process combines human intelligence in the form of expertise-driven algorithms with cloud computing to produce quality results at unprecedented speed and provide first-mover advantage. 2 White paper draft 5/31/2017 Expertise Limitations…Too Many Details Researchers designing multiplex panels need to detect the analytes of interest in a panel, the inclusivity set, while making certain there are no false positives and that undesired analytes are not detected, the exclusivity set. Finding the region where all the inclusivity sequences will be detected but none of the unintended targets is critical to primer design. In some cases, the exclusive set is very close genetically to the inclusive one. Computational power is efficient; in-silico methods can review more possibilities and combinations quicker than the human brain. “The challenge is having the detection of the intended target versus detection of nonintended targets; making sure every time you detect the target you want but only the target you want.” Aude Argillier, Senior Scientist, Design and Assay Development Manchester, Mdx Assay Development, R&D Europe For infectious diseases this is especially challenging. Starting specimens are often blood samples. Blood can contain a variety of infectious agents some which may be co-morbid with the assay’s targets; only the targets of interest must be detected even if other non-desired targets have similarities. An algorithm that locates conserved regions within the inclusive, but not exclusive, set would save substantial time and effort. Although many groups try to do this manually, double-checking a computergenerated response is a more efficient approach. Another challenge is cross reactivity, getting all of the primers in a multiplex design to work as well together as they do alone. Primers must be compatible. Compatible means that the primer sets amplify with similar efficiency and do not form false amplicons. Cross hybridization, heterodimers, homodimers, hair pin formation, secondary structure in general, can all be major blocks for designing large panels and result in an iterative process that drags on. This challenge will only intensify as future panels become larger and more complex. In addition, assay requirements can vary from design to design, further complicating the complexity of the multiplex assay. For example, there are many PCR design parameters, which may affect the efficiency of the PCR process for a particular application, and customization of the process is essential for individual panels. Different multiplex designs may use different chemistries or require different parameters; greater coverage, shorter amplicons or higher melting temperatures may be desirable. Guidelines and good practices exist for parameter selection, and for simple assays this might yield good results. However, for efficient and robust multiplex performance, assuring that only the targets of interest are amplifying, these “best practices” may prove to be too simplistic a tactic, particularly since primer design is the foundation of a high-quality sensitive multiplex assay. Scientists designing assays are subject-matter experts but not necessarily biophysical or thermodynamic experts, and most companies do not have these types of expertise on staff. Scientists can choose all the 3 White paper draft 5/31/2017 correct design parameters to get to a multiplex panel optimization goal but designing the primers to fit those parameters is not straightforward; computer models scale and evaluate primers for complex parameters in a fraction of the time that a human can. Computer algorithms that incorporate the biophysical or thermodynamic expertise to finesse model optimization would be of great benefit, and allow staff to focus their expertise on the subject matter they know best. Use of Sub-Optimal Tools… The Cost of Free Many available freeware tools address certain aspects of multiplex panel design, and many researchers cobble together multiple freeware applications as home-made tools. However, most freeware was not developed for complex multiplex panels but rather simple singleplex design. The limitations of freeware are further exacerbated when researchers try to piece together singleplexes in a large multiplex pool. The cost of free can be iterative cycles of development and testing (Figure X). Figure X. The cost of free can be iterative cycles of development and testing. For example, BLAST (basic local alignment search tool) is often used to make sure only the target of interest is detected. BLAST is meant to determine sequence similarity to infer common evolutionary ancestry and thereby infer function. But sequence similarity does not equal thermodynamic stability. The most common cause of false positives in PCR is the formation of false amplicons due to primer mishybridization, which is not detected by BLAST because it scores hits using sequence similarity rather than using the proper rules for match and mismatch complementarity. DNA Software has well documented that not all mismatches are created equal and they must be accounted for appropriately. The work around is to take the complement of the oligonucleotide then to use BLAST but this produces a number of problems. BLAST provides too many irrelevant hits, the wrong ranking of hits, misses about 4 White paper draft 5/31/2017 80% of thermodynamically-stable hits, does not distinguish extensible from non-extensible hits, does not detect amplicons and cannot do multiplexing. “Combining tools, such as MFOLD, Primer3, Oligo ANALYZER, in just the right way may produce some interesting designs that work fairly well, although they are frequently limited to 3-5 multiplexes. It would be simpler and more time effective to have an integrated application for multiplex design.” Andrew Dunn, Senior Scientist, Preclinical and Clinical Assays, CRISPR Therapeutics Fast High-Quality Results…PanelPlex for Multiplex Primer Design PanelPlex resolves the complexity of multiplex panel primer design. The cloud-based application incorporates proprietary algorithms to address in-depth the thermodynamics underpinnings of PCR and NGS processes, provides consensus designs, and is easy to use for both single and large multiplex assays. PanelPlex uses a degenerate model for primer design for a single target or a family of targets that have a high degree of homology. The end result is the fewest number of primers needed to fit an inclusivity set. Primers and probes against DNA and mRNA are designed in real-time; off-target amplicons are minimized by taking the designs and thermoblasting or vetting them against background genomes. PanelPlex goes through millions of computation in a few hours. The process eliminates iterative trialand-error design cycles, greatly reducing developmental time and cost. The cloud-based application is easy to use, and contains a simplified wizard with only three data-entry screens. The three main steps are to define the exclusivity list, to define the target sequences and design regions or inclusivity set, and to select the hybridization conditions. Exclusivity Set The most popular curated databases, termed “playlists” by DNA Software, are already deposited in a sequence databank in PanelPlex and kept up to date eliminating the need to become an expert in using database curation and formatting software. The option also exists to upload a custom database, or to create a custom database by combining selectable options. Simply input the name, select the type (NCBI entry by accession number, or straight accession) or upload a FASA file. Inclusivity Set The keystone sequence, a highly-related representative of the inclusivity set, is the one most closely related to everyone else in the family. PanelPlex requires the selection of a keystone, which can be provided by accession number or PanelPlex can autodetect and find the best option. When PanelPlex autodetects and designs primers and probes they will have 100% binding to the keystone, thereby improving the answer and providing a better solution more efficiently. 5 White paper draft 5/31/2017 Other options include setting a folding region to define how large an area, such as 250 nucleotides, that PanelPlex will search through in a given time. Detection types, such as SYBR, a molecular beacon or a straight-up probe to convert to TaqMan, can also be selected. PanelPlex will run through a target analysis, align the target with the other inclusivity sequences to find areas of homology, and vet the lists to eliminate cross hybridization. Hybridization Conditions Lastly, select the hybridization condition. Options include default conditions, primer concentration, annealing temperature, reverse transcription temperature, and more. If desired, kits can be selected and proprietary sodium, magnesium, and other additives or denaturants defined. Primer modifications, which will have an effect on binding temperature, can be added on. If a beacon is desired, add the fluorophore and quencher. Hit run. The jobs portal shows job, job status and date stamp. An email is sent when the job is complete. Results show percent coverage; the desired number of solutions can be selected. Drilldown includes primer or probe details, position on keystone target and percent bound, in addition to penalties, such as low complexity score, i.e. runs of A,G,T,or C , biomolecular TM, ΔH and ΔG. More advanced reports and graphics are also available. An electronic laboratory notebook provides the ability to download results and print. PanelPlex provides: Quick Results – Even complex multiplex designs can be completed within 24 hours Economical – Is a fraction of the cost of a full-time employee Expertise – Easy access to “best in class” algorithms, models, computational power and thermodynamic expertise One-Stop Shopping – Eliminates the use of sub-optimal tools Summary PanelPlex addresses the four most frequently encountered problems in multiplex panel design: development time, development cost, expertise limitations and the use of sub-optimal tools. This nextgeneration cloud-based application represents the best of what DNA Software is known for: difficult targets, high-level multiplex, assay optimization and the elimination of trial and error approaches to help companies more efficiently build better diagnostics. PanelPlex is a trademark of DNA Software, Inc. ©2017 DNA Software, Inc. All rights reserved. 6
© Copyright 2026 Paperzz