Capturing Complex Human Genetic Variations using the GS FLX+ System

Transcription

1 SeqCap EZ Library: Technical Note August 2012 Capturing Complex Human Genetic Variations using the GS FLX+ System Sequence Capture of Structural Variants in the Human Genome Primary Authors: Lindsay Freeberg* Mark D Ascenzo* Contributing Authors: Roger Winer *Roche NimbleGen Inc., Madison, WI 454 Life Sciences, a Roche Company, Branford, CT Introduction With the initial sequencing of the human genome, there was great optimism that each gene would soon be matched with a specific genetic disorder or trait. As initial efforts focused primarily on the role of common and rare singlenucleotide polymorphisms (SNPs), it quickly became apparent that achieving this goal would not be simple. One reason was that as more human genomes were being studied, it became evident that structural variation might have more responsibility for genomic diversity in the human population than SNPs. 1,2 Research has associated structural variants (SVs) with a wide variety of diseases and disorders including Crohn s disease, cancer, Parkinson s disease, Alzheimer s disease, epilepsy, autism, and schizophrenia. 2,3 SVs have also been associated with changes in gene dosage. 1 Although the SV role in genetic disease has been clearly established, the effective characterization of SVs has been very difficult because of their complexity, and the fact that they tend to be located in and near repetitive regions of the genome. 1 Many current sequencing platforms combine short read lengths (< 200 bp) with a high number of reads to achieve suitable depth of coverage. Despite the high coverage offered with short read platforms, shorter reads often hinder the mapping and assembly of reads from repetitive or rearranged DNA. 454 Sequencing Systems are the trusted solution to providing high throughput long (>400 bp) sequencing reads. The upgrade to the GS FLX+ System has further extended that by introducing read lengths up to 1000 bp and typical modal read lengths of approximately 700 bp. This technical note describes guidelines for utilizing the long read lengths provided by the GS FLX Titanium Sequencing Kit XL+ and XLR70 with Roche NimbleGen s SeqCap EZ Choice Library. Additionally, the major histocompatibility complex (MHC) region exemplifies the uses of the extra long read lengths for obtaining data in difficult to sequence regions. For life science research only. Not for use in diagnostic procedures. 1

2 Capturing Structural Variants using SeqCap EZ Choice Library and GS FLX+ System Materials and Methods The DNA research samples used in this experiment represent distinct European (CEPH/Utah male) and African (Yoruba female) HapMap study populations (Coriell Institute, Camden, NJ, cat. no. NA12762 and NA19240, respectively). For each HapMap sample, two library types were made. One library type was made following instructions in the NimbleGen SeqCap EZ Library LR User s Guide (v2.0), which will be referred to as XLR70 libraries in this technical note. The second library type was made according to the procedure described in Appendix 1, which will be referred to as XL+ libraries in this technical note. For the XLR70 and XL+ libraries, a different Multiplex Identifier (MID) Barcode Adaptor was used for each sample DNA. A SeqCap EZ Choice Library was developed to capture a large set of known SVs. The design targeted tandem duplications and deletions reported in a recent 1000 Genomes Project paper that used whole genome sequencing from 185 individuals to generate a SV map. 3 In total, the design targeted 1,762 deletion sites and 246 tandem duplication sites to cover a cumulative target region of approximately 4.6 Mbp. For both of the XLR70 and XL+ library types, the NA12762 and NA19240 research samples were captured individually and after mixing in different proportions (50:50, 90:10, and 99:1 for Na12762:NA19240), shown in Table 1. The mixtures were generated to examine the sensitivity of the capture libraries to identify SVs present in sub-mendelian proportions, similar to what might occur after somatic mutation in tumors. The captures were performed in duplicate following instructions in the NimbleGen SeqCap EZ Library LR User s Guide (v2.0) for capture using the XLR70 libraries, and instructions in Appendix 1 for capture using the XL+ libraries. The captures of the two library types will be referred to as XLR70 captures and XL+ captures, respectively. Enrichment for all captures was confirmed via qpcr, which revealed the replicates for each sample library and library type had similar average qpcr enrichment results (data not shown). The replicate with the highest average qpcr was chosen for sequencing. Each capture was loaded onto one region of an eight region PicoTiterPlate (PTP) Device. The XLR70 library captures were sequenced using a GS FLX System and the GS FLX Titanium Sequencing Kit XLR70. The XL+ library captures were sequenced using a GS FLX+ System and the GS FLX Titanium Sequencing Kit XL+. The total number of reads produced for each capture is shown in Table 1. Sequence data from the capture experiments were analyzed using the GS Reference Mapper Software (GS Mapper) provided with the GS FLX+ System.* Capture Overview Capture Number Hybridization Sample Composition XLR70 Captures XL+ Captures XLR70 XL+ NA12762 (%) NA19240 (%) Total Bases Total Reads Modal Read Length Total Bases Total Reads Modal Read Length ,668,225 71, ,895,711 94, ,946, , ,236, , ,615, , ,079, , ,782, , ,273, , ,616, , ,850, , p Table 1: Capture overview. Five types of captures were done with each type of library preparation method. Each capture was done in duplicate. The replicate with the highest average qpcr enrichment estimate was sequenced. * The XLR70 captures were analyzed using GS Mapper version while the XL+ captures were analyzed using GS Mapper version 2.6. Different software versions were used due to the time difference in which the captures were done. The XLR70 capture 2 was reanalyzed by GS Mapper version 2.6 to verify if different software versions would produce different results. A comparison of the two High Confidence Structural Rearrangement files reported by GS Mapper revealed no difference in the location of SVs found. Differences in the GS Mapper Software version should, therefore, not affect the number of SVs found. 2 Capturing Complex Human Genetic Variations using the GS FLX+ System

3 Results The data from the NA19240 captures were first compared between the two library types, XLR70 and XL+, using the High Confidence Structural Rearrangement file reported by GS Mapper. A total of 278 SVs were called with 28 (10%) from the XLR70 capture, 203 (73%) coming from the XL+ capture, and 47 (17%) identified from both libraries (see Figure 1A). While the XL+ capture analysis began with a slightly greater number of reads (6.8%; see Table 1, captures 2 and 7), it is unlikely to account for the much larger discrepancy in sensitivity for SV detection between the two library types. The data suggest that the main factor influencing the identification of SVs is the longer read lengths of the GS FLX+ System, which provide additional sequence context at the targeted sequences and more confidence in identifying SVs. An attempt to verify the SVs detected using public data was performed using a characterization of SVs from HapMap sample NA19240 that was previously completed and made publicly available by Complete Genomics 4 (Mountain View, CA), available at DNAnexus website. Additionally, the whole genome sequence of NA19240 (performed by Complete Genomics 4 ) was also available for the public at the DNAnexus website. Each SV was individually inspected to determine if the location of the SV detected by capture data matched the location of the same SV characterized by Complete Genomics. The Unique Sequence Coverage (coverage of unique and fully mapped reads at each base pair) was used visually to confirm SVs. The visual review was important to confirm SVs, primarily deletions, which were not detected by Complete Genomics SV analysis (see Figure 2 for an example). Percentage Percentage A. B. 100% 75% 50% 25% 0% C. 100% 75% 50% 25% Total Structural Variants (SVs) Common 17% (47) XLR70 Captures 10% (28) XL+ Captures SV Type XLR70 Captures Confirmation of SVs 10 XL+ Captures 73% (203) Inversions Substitution Duplication, Interspersed, Inverted Duplication, Interspersed Insertions Duplication, Tandem Deletions Unsure Unconfirmed Confirmed 0% XL+ Captures XLR70 Captures p Figure 1: Comparison of structural variations (SVs) between the XLR70 Library and XL+ Library NA19240 captures. Panel A shows the distribution of the total SVs detected in the XL+ capture only, XLR70 capture only, and SVs commonly shared. The first number is the percentage of SVs in that category and the second number is the total number of SVs. Panel B shows the distribution of SVs confirmed by the Complete Genomics data (SV and whole genome coverage) in the XL+ capture only and XLR70 capture only. Panel C compares the variety of SVs called by the XL+ capture compared to the XLR70 capture. SeqCap EZ Library: Technical Note 3

4 Capturing Structural Variants using SeqCap EZ Choice Library and GS FLX+ System Next, the NA19240 XLR70 capture and the NA19240 XL+ capture (captures 2 and 7, respectively) were compared. Figure 1B shows the number and percentage of SVs that were confirmed by comparing the XLR70 and XL+ capture data with the data generated by Complete Genomics. SVs that were detected by both the XLR70 and XL+ captures were not included. Since SVs tend to be located in regions of the genome that are difficult to assemble, there was some flexibility given to the endpoints of the SVs when comparing. The flexibility of the endpoints range from a couple of base pairs to base pairs. The discrepancies in the endpoints between the SVs derived from public data and the capture data did not comprise a significant portion of the SV. Therefore, it is unlikely that the flexibility in calling the SV endpoints creates bias in the data. SVs that were confirmed were usually detected as an SV by both capture and Complete Genomics, and/or had a significant change in the whole genome sequencing coverage depth (e.g. greatly reduced coverage at a deletion). The majority of the SVs were confirmed for both types of capture: 64% for XLR70 capture and 68% for XL+ capture. SVs that were unconfirmed (36% for XLR70 capture and 21% for XL+capture) were usually not detected by Complete Genomics and did not have a significant change in the whole genome sequencing coverage depth. It is possible that these unconfirmed SVs are still novel SVs that are difficult to confirm using the methods stated in this technical note, such as duplications, inversions, and insertions. In addition, the SVs could be heterozygous in the DNA sample, which would make it difficult to detect and verify using whole genome sequencing coverage depth. There were a few (10%) detected SVs that were neither confirmed nor unconfirmed in the XL+ capture, which was usually due to not being detected as an SV by Complete Genomics and having a shift in whole genome coverage depth that could not be easily distinguished from the normal variance of coverage depth throughout the genome. A closer examination of all the SVs detected revealed that more types of SVs were called in the XL+ captures (see Figure 1C). All SVs detected were used in this analysis since it is possible that the unconfirmed SVs are still novel SVs that are difficult to confirm using the methods in this technical note. The majority of SVs called by both captures were deletions. In addition, the XL+ captures also identified insertions, several types of duplications, and a substitution. The increased read length in the XL+ captures provides data over a greater portion of the SVs, which could allow insertions, duplications, etc., to be more accurately mapped and called which increases confidence in making SV calls. To determine if SVs could be detected when present at sub- Mendelian frequencies, the SVs from the XLR70 captures were compared to find a variant that was present in captures containing HapMap sample NA19240 (captures 2-5) and absent in HapMap sample NA12762 (see Table 1, capture 1). One variant, a 324 bp deletion at chromosome 5 position 172,632,853 was detected in captures 2-4, down to a dilution containing only 10% DNA from sample NA The variant was not detected in the dilution containing 1% DNA from sample NA19240 nor in sample NA12762 from the XLR70 captures (captures 5 and 1, respectively). To check if the variant was observed but did not meet the automated variant calling thresholds for the 1% DNA dilution of NA19240 (capture 5), the All Structural Rearrangement file reported by GS Mapper was searched for the variant. The variant was not present in the All Structural Rearrangement file. Additionally, XLR70 capture 5 was sequenced at 454 Life Sciences on a full PTP Device. At this depth, the variant was detected in XLR70 capture 5. Examination of the XL+ capture data revealed that the same deletion was detected in captures 7-10, down to the dilution containing 1% DNA from sample NA As with the XLR70 capture data, this variant was not detected in sample NA12762 XL+ capture data (capture 6). It is unclear why the variant was detected in the XL+ captures and not the XLR70 captures when sequenced at a depth of one region in an 8-region PTP Device. It could be due to the longer read length which increased coverage in the region or it could be due to random sampling effects. Conclusions By targeting areas where SVs were believed to occur, it was possible to detect them using Sequence Capture. The majority of the SVs identified using both the XLR70 and XL+ captures were confirmed using a combination of two analysis approaches described above. Using a similar number of reads, experiments capturing from the XL+ captures identified a considerably larger number and variety of SVs than experiments capturing from XLR70 captures, highlighting a valuable application for Sequence Capture paired with extra-long sequencing technology provided by the GS FLX+ System. 4 Capturing Complex Human Genetic Variations using the GS FLX+ System

5 A. Signal Map Software Design Files Probes Deletion Region SVs CNVs Indels B. DNAnexus Software Repetitive Elements Whole Genome Coverage SVs p Figure 2: A 4.4 kb deletion at position 51,561,806 of chromosome 7. Panel A shows data from SignalMap Software (Roche NimbleGen) and Panel B is data from DNAnexus Software. In Panel A, the top 3 tracks are from the design file. The 4 th track (gold bars) shows the location of the probes with the height representing the number of probe bases at that location. The 5 th track (green box) is the location of the deletion which spans from position 51,561,806 to 51,566,184. The last three tracks in Panel A separately indicate the SVs, CNVs, and Indels detected by Complete Genomics 4. The SVs track shows that no SVs were detected by Complete Genomics. The CNV track also appears to not confirm the deletion since it shows a drop from 1.2 to 0.6 near the start of the deletion, but it does not increase near the end of the deletion. In Panel B, the first track denotes the location of repetitive elements. The middle track is the coverage of whole genome sequencing by Complete Genomics, and the last track displays SVs detected by Complete Genomics. The Complete Genomics SV track does not indicate a detected SV, just as in the SignalMap Software SV track. However, when comparing the positioning of the deletion to the whole genome sequencing coverage data, there appears to be reduced coverage which could indicate an uncharacterized SV. The deletion was detected using the XL+ capture 7 and the GS FLX+ Sequencing System. SeqCap EZ Library: Technical Note 5

6 Capturing the MHC region using GS FLX+ Series XL+ Libraries The major histocompatibility complex (MHC) is one of the most studied regions in the human genome because of its association with autoimmune, infectious, and inflammatory diseases. It plays a important role in organ transplant rejection as it codes for cell surface proteins that are responsible for differentiating between host and foreign cells. The MHC region contains a high occurrence of polymorphisms, and there are 7,269 allele sequences (MHC Database5, 6 v3.7) for this locus according to the MHC database. In addition, the MHC region is known for being a highly variable and repetitive region with low complexity, which makes it difficult to sequence and assemble. However, the use of longer sequencing reads, such as those obtained from the 454 GS FLX+ System, can make it easier to sequence and assemble this region. Material and Methods The capture used in these experiments was designed to enrich for the MHC region (3.6 Mb) of chromosome 6 and additional content flanking the MHC region. 7 In total, these regions represent approximately 4.8 Mb of chromosmome 6 and include additional targeted content present in the HG19 haplotype reference sequences (APD_HAP1, COX_HAP2, DBB_HAP3, MANN_HAP4, MCF_HAP5, QBL_HAP6, SSTO_HAP7). (Note: The design used in this technical note and a second generation design will be available for ordering from Roche NimbleGen.) The XL+ libraries were prepared using gdna from two HapMap subjects: NS12911 (Coriell Institute, Camden, NJ cat. no. NS12911) and UCLA 590 from the International UCLA HLA DNA Exchange (University of California, Los Angeles, CA), respectively. For each gdna sample, the libraries were captured in duplicate. The average qpcr estimates of enrichment were approximately equivalent between replicates for each sample library (data not shown). All captured samples were sequenced using the GS FLX+ Sequencer at 454 Life Sciences. Each captured sample was loaded onto four regions of an eight region PTP Device. Reads generated from the capture experiments for the captured samples were analyzed using the GS Reference Mapper Software (GS Mapper) version The duplicate captures of each gdna sample were pooled together. The GS Mapper was used to map reads against the HG19 reference genome with haplotype reference sequences excluded. Results Sequencing and Coverage Statistics Sequencing generated 1,377,097 and 1,441,618 number of reads for samples NA12911 and UCLA 509, respectively. For sequenced read length, gdna sample NS12911 had an average read length of 653 whereas gdna sample UCLA 509 had an average read length of 637. Sequence coverage results over the MHC capture region are shown in Table 2. Using the long-read chemistry of the GS FLX+ Instrument, it was possible to cover greater than 95% of the primary target region for both samples. Sequence Capture probes cannot be placed across the entire MHC region due the repetitive nature of some positions within the region (Figure 3 and Figure 4). These regions include low complexity elements such as SINEs and LINEs. However, using the longer reads produced by the GS FLX+ System, it is possible to read into and cover regions where probes have not been placed, even if the probes are hundreds of base pairs apart (Figure 4). MHC region coverage Kit XL+ Sample 590 NS12911 Primary Target 4,754,830 4,754,830 Primary Target Coverage 4,562,280 4,554,758 Primary Target Coverage (%) 96.0% 95.8% Capture Target 3,480,248 3,480,248 Capture Target Coverage 3,430,168 3,429,435 Capture Target Coverage (%) 98.56% 98.54% p Table 2: Sequence coverage of MHC region. Conclusions Regions such as the MHC contain highly variable sequences with low complexity, which makes sequencing and assembly difficult. With Sequence Capture, it is possible to target these regions to increase assembly accuracy by creating probes specific for these regions. Unfortunately, there are some regions of low complexity where probes cannot be placed. However, by using longer reads, such as those produced by the GS FLX+ System, it is possible to cover gaps between regions that are hundreds of base pairs apart. The synergy of Sequence Capture with the GS FLX+ System can make it possible to fully sequence and assemble these regions. 6 Capturing Complex Human Genetic Variations using the GS FLX+ System

7 p Figure 3: Sequence coverage over MHC region. The first track, 16903_primary, represents the sequence coverage for gdna captured sample 509. The second track, 16905_primary, represents the sequence coverage for gdna captured sample NS The third track is the captured target area while the last track is the primary target region. Image was generated using SignalMap, version (Roche NimbleGen). p Figure 4: Sequence coverage over low probe density regions. Sequence coverage spanning regions where probes cannot be placed due to repetitive and low complexity regions. Despite being separated by hundreds of base pairs, the regions between the probes are able to be covered due to the long reads of the GS FLX+ System. The first track, 16903_primary, represents the sequence coverage for gdna captured sample 509. The second track, 16905_primary, represents the sequence coverage for gdna captured sample NS The third track is the captured target area while the last track is the primary target region. Image was generated using SignalMap, version (Roche NimbleGen). SeqCap EZ Library: Technical Note 7

8 Summary By using the SeqCap EZ Choice Library to target these difficult areas and the extra long reads from the GS FLX+ System, it is possible to explore these areas more thoroughly than before. Even the read lengths provided by the XLR70 Sequencing Kit are not sufficient to fully characterize many of these complex variations as is demonstrated by the substantial increase in SV detection provided by the GS FLX+ System with the XL+ Sequencing Kit. One of the issues for any sequence capture platform is the challenge of placing probes in repetitive and low complexity regions. Such regions are common in the MHC locus. However, by combining the longer capture lengths with the longer reads, it is possible to sequence the regions between probes thus providing a much more comprehensive characterization of the most highly variable portions of the human genome where much complex biology resides. 8 Capturing Complex Human Genetic Variations using the GS FLX+ System

9 References 1. Alkan, C. et al. Genome structural variation discovery and genotyping. Nature Reviews Genetics (2011) 12: Stankiewicz, P. and J.R. Lupski. Structural Variation in the Human Genome and its Role in Disease. Annu. Rev. Med. (2010) 61: Mills, R.E. et al. Mapping copy number variation by population-scale genome sequencing. Nature (2011) 470: Drmanac, R. et al. Human Genome Sequencing Using Unchained Base Reads on Self-Assembling DNA Nanoarrays. Science (2010) 327: Robinson, J. et al. The IMGT/HLA Database. Nucleic Acids Research (2011) 39 Suppl 1:D Robinson, J. et al., IMGT/HLA a sequence database for the human major histocompatibility complex. Tissue Antigens (2000) 55: The MHC sequencing consortium. Complete sequence and gene map of a human major histocompatibility complex. Nature (1999) 401: SeqCap EZ Library: Technical Note 9

10 Appendix 1: Instructions for SeqCap EZ Choice Library Sequence Capture using the XL+ Library Reagents, Consumables & Equipment Refer to NimbleGen SeqCap EZ Library LR User s Guide and 454 Life Sciences Rapid Library Preparation Method Manual for GS FLX+ Series - XL+ Reagents, Consumables, and Equipment. There are no additional reagents, consumables, or equipment needed. Note: The Rapid Library Preparation XL+ requires a 1 µg input of gdna. References 454 Life Sciences Rapid Library Preparation Method Manual GS FLX+ Series - XL+ (May 2011) NimbleGen SeqCap EZ Library LR User s Guide (v2.0) Invitrogen Quant-iT PicoGreen dsdna Assay Kit Manual (June ) Pre-Library Preparation 1. Review the Rapid Library Preparation Method Manual GS FLX+ Series - XL+ for Equipment, Labware, and Consumables. 2. See Chapter 1. Before You Begin of the NimbleGen SeqCap EZ Library LR User s Guide to become familiar with the EZ Choice product for Equipment, Labware, and Consumables necessary for capture. 3. See Chapter 2. Storage of the SeqCap EZ Library of the NimbleGen SeqCap EZ Library LR User s Guide for instruction on storing the EZ Choice Library. XL+ Sample Library Preparation and QC 1. For DNA sample quality requirements, see Section 2 Sample Requirements of the XL+ Rapid Library Preparation Method Manual. 2. To prepare XL+ libraries, refer to Section 3.1 DNA Fragmentation by Nebulization through Section 3.5 Small Fragment Removal of the of Rapid Library Preparation Method Manual with the following exceptions: When eluting the sample library from the AMPure Beads (Step 14 of Section 3.5), use PCR grade water. The EDTA present in TE will chelate the Mg 2 + and inhibit the Pre-Capture LM-PCR. Note: The sample library is more susceptible to degradation because it is eluted in water. Therefore, it is recommended to proceed to the Section 3 Pre-Capture LM-PCR Through qpcr of this Appendix as soon as possible. 3. For library quality assessment and quantification, refer to sections 3.6 Library Quantitation, 3.7 Library Quality Assessment, and section 4 Appendix of Rapid Library Preparation Method Manual with the following recommendation: Use PCR grade water instead of TE to dilute the standard (Section 3.6.1). 10 Capturing Complex Human Genetic Variations using the GS FLX+ System

11 Pre-Capture LM-PCR Through qpcr 1. Follow Chapter 4. Amplifying the Sample Library Using LM-PCR through Chapter 8. Measuring Enrichment Using qpcr of NimbleGen SeqCap EZ Library LR User s Guide on how to proceed for the rest of the SeqCap Process. For both Pre- and Post-Capture LM-PCR, it was concluded that no additional cycles or increased extension times are necessary to amplify XL+ Rapid libraries. Notes: The fragment size distribution might shift to a lower size distribution after Pre-Capture LM-PCR. However, the size ranges for LM-PCR Bioanalyzer should be similar to the table in Step 2 of Section 3.7 Library Quality Assessment and Figure 4 of Appendix in the XL+ Rapid Library Preparation Method Manual. Although Pre- and Post-Capture LM-PCR fragments sizes are up to 3 Kb, sequencing up to this length may not be possible due to a limitation related to the GS FLX+ System. 2. For multiplexing, see Appendix B. Instructions for Multiplex Sequence Capture with NimbleGen SeqCap EZ Libraries of NimbleGen SeqCap EZ Library LR User s Guide for instruction. empcr Amplification and Sequencing 1. Follow Chapter 9. Preparing empcr Amplification and Sequencing of NimbleGen SeqCap EZ Library LR User s Guide on instructions for post-capture guidelines for sequencing preparations with the following exception: For dilution and storage, refer to section 3.8 Preparing Working Aliquots of the Rapid Library Preparation Method Manual. Refer to the proper 454 GS FLX+ Series XL+ empcr Amplification and Sequencing Manuals for how to proceed for empcr Amplification and Sequencing. Notes: At the time of publication, the XL+ Rapid Libraries are not supported on the GS Junior Instrument. SeqCap EZ Library: Technical Note 11

12 Ordering Information Product Cat. No. Pack Size GS FLX+ Instrument Instrument GS FLX Titanium Sequencing Kit XLR kit (for one sequencing run) GS FLX Titanium Sequencing Kit XL kit (for one sequencing run) GS FLX Titanium Rapid Library Preparation Kit kit (for 12 library preparations) SeqCap EZ Choice Library reactions reactions reactions reactions reactions reactions SeqCap EZ Hybridization and Wash Kit reactions reactions RESTRICTION ON USE: Purchaser is only authorized to use 454 Sequencing System Instruments with PicoTiterPlate devices supplied by 454 Life Sciences Corporation and in conformity with the operating procedures contained in the 454 Sequencing System manuals and guides. This product incorporates technology licensed from Promega Corporation that is protected under US Patents 5,583,024 (exp. 12/10/2013), 5,674,713 (exp. 10/7/2014) and 5,700,673 (exp. 12/23/2014). Restriction on use. As a condition of sale of this product, purchaser agrees not to use the product to perform less than 1,536 sequencing reactions on a sample or samples without changing the substrate. Failure to comply with this restriction will result in an infringement of patent rights and other intellectual property rights of seller or third parties and a breach of the terms of sale of this product. NOTICE: This product may be subject to certain use restrictions. Before using this product please refer to the Online Technical Support page ( and search under the product number or the product name, whether this product is subject to a license disclaimer containing use restrictions. For life science research only. Not for use in diagnostic procedures. 454, 454 LIFE SCIENCES, 454 SEQUENCING, EMPCR, GS FLX, GS FLX TITANIUM, GS JUNIOR, NIMBLEGEN, PICOTITERPLATE, PTP, and SEQCAP are trademarks of Roche. All other product names and trademarks are the property of their respective owners. Published by: Roche Diagnostics GmbH Sandhofer Straße Mannheim Germany 2012 Roche NimbleGen, Inc. All rights reserved /12