Introduction to High Resolution Melt Analysis
- 1. HRM Analysis Francisco Bizouarn International Field Application Specialist Gene Expression Division
- 2. Overview • Introduction to High Resolution Melt (HRM) • Applications • Important Considerations • Assay Design and Optimization • Precision Melt Analysis software
- 3. High Resolution Melt • Post-PCR melt analysis method • Discriminates dsDNA based on sequence length, GC content or strand complementarity • Detects a single base difference • Rapid, inexpensive sequence screening method – Mutation sequence can be unknown – Samples are further processed to identify mutation sequence • Increased specificity and sensitivity Sample PCR HRM
- 4. HRM Applications • Mutation discovery/gene scanning • SNP genotyping 95% of all applications • DNA methylation analysis • Species identification • DNA fingerprinting • Screening for loss of heterozygosity • Allelic prevalence in a population • Characterization of haplotype blocks • HLA compatibility testing • Identification of candidate predisposition genes
- 5. Melt Curve Analysis • After real-time PCR amplification, a melt curve is performed in presence of a DNA binding “saturation dye” • Melting temperature (Tm) – DNA is half double and half single-stranded – Depends on nucleotide content and length Double Single Stranded DNA Stranded Tm
- 6. Melt Curve Analysis • Distinguish products based on their Tms – Plot negative rate of change of fluorescence vs. temp (-dI/dT) for easy discrimination of products based on their Tms
- 7. Applications
- 8. SNP Genotyping • A single base substitution, prevalent to 1% in a population • Use HRM analysis to identify samples containing known single nucleotide polymorphisms • Not all SNPs are equally easy to differentiate SNP Class Base Change Rarity (in human genome) 1 C/T and G/A 64% 2 C/A and G/T 20% 3 C/G 9% 4 A/T 7% SNP classes defined by Venter et al (2001)
- 9. HRM Genotyping Curves • The observed fluorescence curve shown in black is the composite of all four possible duplexes C G T A C ^ v A T ^ v G
- 10. Mendelian Genetics Review Homozygote WT Heterozygote Homozygote mutant C C T C T T Allele 1 C C T G G A Allele 2 C T T G A A C Melting C G T G T A A C Duplexes C T T G A A G
- 11. Normalization • Pre-melt (initial) and post-melt (final) fluorescence signals of all samples are normalized to relative values of 100% and 0% • Eliminates differences in background fluorescence between curves
- 12. Difference Plot • Magnify curve differences by subtracting each curve from the most abundant type or from a user-defined reference • Sets a baseline, so small differences become visible
- 13. Cluster Analysis • The software clusters similar curve shapes automatically into groups representing different genotypes (sequences) • The software then auto-calls samples to a genotype depending on where their curve shape clusters C/T C/C T/T
- 14. Class 1 SNP Mutation G>A • Hemochromatosis (HFE), C282Y mutation • 75bp amplicon • Genomic DNA from human blood, using SsoFAST Eva Green Mix • Melt Study results from 3 melt files (12 samples per genotype) G/G A/A G/A
- 15. Class 3 SNP Mutation C>G • Hemochromatosis (HFE) gene, H63D mutation • 100bp amplicon • Genomic DNA from human blood, using SsoFAST Eva Green Mix • 10 samples of each genotype C/C G/G C/G
- 16. Class 4 SNP Mutation A>T • Hemochromatosis (HFE) gene, S65C mutation • 100bp amplicon • Genomic DNA from human blood, using SsoFAST Eva Green Mix • 3 samples of each genotypes
- 18. Methylation • 2-5% of the cytosines in the genome are methylated • 5th base • Epigenetic information - may change over time • No effect on base pairing Mtase=methyltransferase Mtase S-Adenosyl Methionine SAH SAM
- 19. Methylation Sites • Only cytosine preceding guanines are methylation sites- CpG • CpG- dinucleotides are unevenly distributed in the genome • CpG-islands 1000-2000 bases long with 10-20 times higher CpG content • Promoter regions and first exon of many protein coding genes
- 20. Methylation and PCR • Taq polymerase does not distinguish between cytosine and 5- methylcytosine • After the first PCR cycles, all 5-,methyl-cytosine will be substituted for non-methylated cytosines • PCR on untreated methylated DNA will erase the epigenetic information
- 21. Traditional Approach: Bisulfite Treatment • Converts non-methylated Cytosines to Uracil • Methylated Cytosines remain intact • During PCR the Uracils are replaced by Thymine • GC pairs are shifted to AT at non-methylated CpGs and non CpG Cytosines resulting in a Tm change
- 22. Methylation Data Analysis • Bisulfite treated DNA with all CpG-sites methylated will have higher Tm than if non-methylated • C:G vs A:T • Bisulfite treated samples where: Higher Tm in samples with promoter methylated in all cells compared to samples with promoter methylated in only 30% of the cells
- 23. Methylation Assay • CDH1 (Cadherin E), • 113bp amplicon • iQ SYBR Green Supermix • Methylated gDNA diluted with unmethylated gDNA
- 24. HRM – Species identification Mycoplasma species identification with SsoFast EvaGreen Supermix and Precision Melt Analysis Software • Samples amplified directly from tissue culture supernatants • rpoB gene amplified (400 to 600 bp products) • 4 known and 10 unknown species M. ureolyticum M. argninig M. pirum M. fermentans M. hyorhinis M. hominis M. orale M. gallisepticum M. salivarium M. laidlawii
- 25. Mutation Discovery • aka Mutation screening, SNP discovery, DNA re-sequencing, gene scanning • Rapidly screen many samples to identify the few that have mutations • HRM gene scanning can rapidly identify samples with a mutation, however, further analysis is still required to identify the mutation – DNA sequencing – Compare against known genotype profiles using HRM – Perform HRM DNA Matching experiments
- 26. DNA Matching • aka species identification, compatible donor screening • Quickly identify similar sequences – forensic sibling identification, screening for pathogenic or antibiotic-resistant bacterial species etc • Mix samples together and then perform HRM • Downstream Analysis Still Required • Not all homozygotes can be distinguished • May require heteroduplex analysis where specific rations of known genotypes are added to unknown samples and melted
- 27. Genotyping for Sequence Insertions/Deletions/Repeats • Looking for trinucleotide repeats • 4 possible genotypes • Fig. C. shows all possible duplexes formed from the 4 genotypes
- 29. Amplicon Melting • HRM assays are comparisons of dissociation patterns. – If amplicons dissociated a a specific temperature we would see the following. ACGTACGTACGTACGTACGTACGTACGTACGTACGTACGTACGT TGCATGCATGCATGCATGCATGCATGCATGCATGCATGCATGCA – Amplicons dissociate at a rate that will vary according to sequence homology, salt, length, etc… AATTAAT TAATTAATTA ACGTACGTACGTACGTACGTACGTACGTACGTACGTACGTACGT TGCATGCATGCATGCATGCATGCATGCATGCATGCATGCATGCA TTAATTA ATTAATTAAT
- 30. Mendelian Genetics Review Homozygote WT Heterozygote Homozygote mutant C C T C T T Allele 1 C C T G G A Allele 2 C T T G A A C Melting C G T G T A A C Duplexes C T T G A A G
- 31. Melting Profiles Homo WT Homo Mut Hetero
- 32. Small differences Homozygotes produce a single re-associated products A T A T Heterozygotes produce a mixed population of re-associated products T A T T T A A A
- 33. Small differences • Heterozygotes can easily be detected. A T A T T A A T
- 34. Amazingly • Very small differences can be detected. – The juxtaposition of bases A T A T T A T A
- 35. What’s really going on? Max or Total Signal Baseline Noise Level (dye, DNA, instrument) Change of florescence due to temperature effect (buffering and pH) Change of florescence resulting from the melting of double stranded DNA
- 36. DNA melting phase • Comprised of; – Continued change in florescent signal due to buffering effect. – Dissociation of amplicons that are the desired amplification product – Dissociation of amplicons that are not the desired amplification product. – Dissociation of template DNA (to a lesser extent)
- 37. Mechanics of High Resolution Melt What are HRM profiles? • The quick version – The dissociation curves of dye bound PCR amplification products are double baselined and subsequently compared to one another for differences in their melting profiles. • The in depth version – Signal vs noise ratio difference analysis between samples at varying temperatures.
- 38. How is differentiation done • These fluorescent results are plotted with signal intensity on the “Y” axis. • Data for the various temperature increments is plotted sequentially on the “X” axis. • Data is then double baselined using values before and after the melt phase. • Data is re-scaled (normalized) such that each profile ranges from 0 to 1 in in fluorescent intensity.
- 39. How is differentiation done • The differentiation process is repeated for each temperature point at which the data was collected. • The larger the difference in fluorescent readings at a specific temperature, the larger the difference in in the “Difference RFU” graph
- 40. How is differentiation done • Subsequently, the smaller the difference in fluorescent readings at a specific temperature, the smaller the difference in in the “Difference RFU” graph. • Large fluorescent differences make analysis simpler.
- 41. Amplicon Size • HRM assays are comparisons of dissociation patterns – It stands to reason that the larger the signal difference between samples, the easier it is to differentiate and identify them. – As such, smaller amplicons generally work better than larger ones. ACGTACGTACGTACGTACGTAC A GTACGTACGTACGTACGTACGT TGCATGCATGCATGCATGCATG T CATGCATGCATGCATGCATGCA ACGTACGTACGTACGTACGTAC G GTACGTACGTACGTACGTACGT TGCATGCATGCATGCATGCATG C CATGCATGCATGCATGCATGCA – The larger the amplicon the lower the over signal difference as the signal differences is partially obscured by the overall noise. ACGTACGTACGTACGTACGTACACGTACGTACGTACGTACGTAC A GTACGTACGTACGTACGTACGTACGTACGTACGTACGTACGTAC TGCATGCATGCATGCATGCATGTGCATGCATGCATGCATGCATG T CATGCATGCATGCATGCATGCATGCATGCATGCATGCATGCATG ACGTACGTACGTACGTACGTACACGTACGTACGTACGTACGTAC G GTACGTACGTACGTACGTACGTACGTACGTACGTACGTACGTAC TGCATGCATGCATGCATGCATGTGCATGCATGCATGCATGCATG C CATGCATGCATGCATGCATGCATGCATGCATGCATGCATGCATG
- 42. Amplicon Size 50 bp amplicon • Comparison of A/A and T/T heterozygotes • 300nM primer • Optimal annealing temp • Clustering and percent confidence are increased 500 bp amplicon using smaller amplicons
- 43. 96 well A/A A/T T/T assay • The most difficult SNP class – 50pb amplicon – Auto analysis – One failed wells out of 96 – Rest all accurately clustered
- 44. Melting Resolution • This widely distributed table leads to some confusion. SNP Class Base Change Typical Tm Shift Rarity (in human genome) 1 C/T and G/A Large (>0.5oC) 64% 2 C/A and G/T 20% 3 C/G 9% 4 A/T Very Small (<0.2oC) 7% SNP classes defined by Venter et al (2001) What does this column really mean?
- 45. Resolution increments necessary for HRM • It is often assumed that to perform “High Resolution Melt” it is necessary to take readings across a temperature range using very small increments. • More important than the ability to perform thin “slices” is the capability to detect small differences. • Although smaller increments than those used in classic qPCR assays should be used, depending on the assay increments of 0.5 degrees can be used. • Critical parameters to successful HRM assays include proper assay design, good analysis software, clean template and uniformity in assay setup.
- 46. Class 4 SNP 50pb amplicon Melt inc 0.1oC Melt inc 0.2oC
- 47. Class 4 SNP 50pb amplicon Melt inc 0.3oC Melt inc 0.4oC
- 48. Class 4 SNP 50pb amplicon Melt inc 0.5oC • SNP detection can be performed with increments of 0.5 degrees (max tested in this set of assays) when using small amplicons even with class 4 SNP’s. • Heterozygote samples are the easiest to spot and generally give a very high confidence level when auto clustering.
- 49. Class 3 SNP 100 pb amplicon Melt inc 0.1oC Melt inc 0.2oC
- 50. Class 3 SNP 100 pb amplicon Melt inc 0.3oC Melt inc 0.4oC
- 51. Class 3 SNP 100 pb amplicon Melt inc 0.5oC • Simplicity of the SNP is also a factor in melt analysis. • Here a class 3 SNP is easily clustered at 0.5 degree increments.
- 52. Assay design and optimization
- 53. Assay Design Workflow Identify Sequence Design Primers Evaluate Assay NCBI Beacon Designer, Evaluate real-time PCR Sequence information Primer3, BLAST, and HRM assays MFOLD
- 54. Target Sequence • SNP analysis – Identify the right sequence – Search for SNPs based on gene, location or function – Find variation sites (avoid variations that impact melt curves) • Methylation Assays – CpG sites in primers and within the sequence – Fragment length • NCBI SNP Databases – http://www.ncbi.nlm.nih.gov/SNP
- 55. Primer Design • Primer guidelines – BLAST primer sequences – 18-24 bases – 40-60% G/C – Balanced distribution of G/C and A/T bases – Annealing between at 55-65 C – No internal secondary structures (hair-pins) • Primer pairs – Similar Tms, within 2-3 C – No significant complementarity (> 2-3bp), especially in 3’ ends • Primer binding sites – Avoid targets with secondary structure – Avoid pseudogenes
- 56. Experiment Considerations • Amplify a single product at high efficiency – No primer-dimers or non-specific products • Generate sufficient PCR product (C(t)s 30) • Samples need equal PCR efficiencies and plateau fluorescence for comparison • Analyze short PCR products, the smaller the better • Uniform reaction mix/sample concentrations • Capture data over at least a 10 C melt curve range
- 57. Amplicon Design • Use similar criteria as for SYBR Green assay • Short amplicons maximize differences in melting behavior between similar sequences – 70-150bp is desired (50-250bp acceptable) – Longer amplicons yield more complex profiles, with multiple melting domains, consider melt domain complexity (M-fold) – Avoid areas of secondary structure and high GC content – Local sequence context can influence mutation detection – Overall GC content and position of the mutation in the fragment does not have a significant effect on mutation detection • Determine folding characteristics at annealing temperature (DINAMelt)
- 58. PCR Evaluation • Poor PCR optimization = Poor HRM resolution • No primer-dimers or non-specific products – Thermal gradient optimization of annealing temperature – Increase annealing temperature/decrease MgCl2 concentration to increase specificity – Run No Template Controls (NTC) – Optimize primer concentrations (100nM steps) • Efficient PCR amplification, want similar plateau fluorescence • Generate sufficient product, C(t) values below 30 – Degraded material or too little material, increase concentration • Do not use UNG enzymes for methylation assays Always check amplification curves prior to HRM analysis
- 59. PCR Optimization • Starting material is key to good results – Handle samples properly prior to analysis – Use a consistent amount of starting material • Amplification – Use a Hot-Start, high-fidelity enzyme – Shorten protocol steps – Optional: hold at 72 C after amplification – 95 C to denature PCR products • Hold at 40-50 C for heteroduplex formation
- 60. HRM Troubleshooting • Problems with interpretation of HRM results – Include controls for each known variant in the test population – Make fresh reaction mixes for each new run – Run plates within two hours of preparation – Perform melt curve immediately following amplification • Inconsistent melt behavior can also occur due to variations in: – MgCl2 and buffer salts – Taq storage buffer additives – Intercalating dye type and concentration – Reaction volumes – Melt ramp rate
- 61. HRM Assays Summary • Consistency in reaction setup and reagent use is necessary for comparisons of samples • Protocols vary depending on the application • Results vary depending on DNA template quality and the sequence • Avoid primer-dimers and additional products that effect melting behavior • Use negative controls and standards • Any ambiguous samples should be sequenced
- 62. Precision Melt Analysis software
- 63. Workflow 1. Set up reactions using HRM compatible reagents SsoFAST Eva Green Supermix 2. Run amp + high resolution melt protocol on CFX96 or CFX384 98°C for 2 min 98°C for 2-5s 55-60°C for 2-10sec (plate read) Go to step 2, 39 more times 98°C for 1 min 70°C for 1 min Melt curve 75°C to 95°C increment 0.2°C 10 sec hold (plate read) 3. In Precision Melt Analysis software, import the data file (.pcrd) to create a melt file (.melt)
- 64. Data Analysis • Multiple views of data, with easy interpretation of results • Analyze multiple experiments from a single plate using the Well Group feature
- 65. Charts View • Quickly compare amplification plots and melt curves
- 66. Percent Confidence • Provides a percentage chance that a given well is correctly categorized within the assigned cluster • It is based on the number of standard deviations the sample is from the mean of the cluster. This assumes that the found "cluster means and standard deviations" are accurate descriptions of the real probability distributions of the data