ChIP-seq has emerged as the favored method for producing genome-wide epigenetic profiles for chromatin proteins and transcription factors, yet sample preparation for deep sequencing still remains a hurdle due to the limited quantity and quality of the DNA sample following de-crosslinking. When only a few cells are available, poor quality or over-amplified ChIP libraries may require additional replicates and sequencing to obtain confident peak calls. Swift Biosciences has developed a sample preparation solution to overcome these limitations. The Accel-NGS® 2S Plus DNA Library Kit utilizes as little as 100 pg of starting material to generate highly complex ChIP-seq libraries for NGS and reduce the sequencing depth requirements.
SWIFT PRODUCT LINES COMPATIBLE WITH ChIP-SEQ:
The Swift advantage:
- As little as 100 pg input required to create libraries from low yield IPs.
- Two-hour single tube “with bead” protocol will help you process more samples faster.
- Superior library complexity increases confidence in peak calls.
- Even coverage of AT-/GC-rich regions maximizes coverage of difficult targets.
Superior ChIP-Seq Data from 100 pg
A comparison of seven different library preparation kits by researchers at Oslo University Hospital sought to determine the method which produced the best ChIP-Seq data (see publication here). Tri-methylated H3K4 ChIP DNA was used as input at both 1 ng and 0.1 ng quantities. Sequencing data from these libraries was compared to a PCR-free library constructed from 100 ng of the same ChIP DNA. The Accel-NGS 2S PCR-Free Kit was used to construct the PCR-free libraries, as it was found to be the only library preparation kit capable of constructing PCR-free libraries from 100 ng of input DNA.
Experimental Design of H3K4me3 ChIP-Seq Comparison
Seven library preparation methods were evaluated with 1 ng and 0.1 ng inputs, and compared to PCR-free libraries. An Accel-NGS 2S PCR-Free Kit was used as it was the only library kit tested capable of creating a PCR-free library from 100 ng of DNA.
Sequencing data was analyzed to compare the number of uniquely mapped reads from each library. At 1 ng inputs, the Accel-NGS 2S Plus DNA Library Kit produced libraries with the most uniquely mapped reads, comparable to the PCR-free libraries. At 0.1 ng inputs, the uniquely mapped reads from the Accel-NGS 2S Plus libraries remained high, while all other competitor methods lost a large portion of uniquely mapped reads.
Accel-NGS 2S Plus Libraries Maximize Uniquely Mapped Reads
Uniquely mapped reads from each of the library preparation methods is plotted compared to the 100 ng PCR-free library at the top (dark blue). The Accel-NGS 2S Plus libraries maintained the highest number of uniquely mapped reads for both 1 ng and 0.1 ng inputs.
Histone ChIP-Seq from as Little as 1,000 Cells
The Accel-NGS 2S Plus Kit allows you to push the lower limits of input for your ChIP-seq applications. Using 1,000 GM12878 lymphoblast cells, a ChIP was performed for histone methylation (H3K27me3), followed by library preparation with the Accel-NGS 2S Plus Kit. Sequencing data from this library illustrates the comparability of its peaks to 2S Plus data from 5,000 lymphoblast cells and ENCODE sequencing data from 20 million cells. These data indicate that the quality of peak calling with the Accel-NGS 2S Plus Kit is preserved even with limiting samples.
Histone Methylation ChIP-Seq from 1,000 Cells
GM12878 lymphoblast cells were used to perform ChIP for H3K27me3, followed by library preparation with the Accel-NGS 2S Plus Kit. ENCODE data using 20 million cells is shown for comparison. Data generated by Active Motif.
Stronger, Sharper Peaks Than Traditional Library Prep
The performance of the Swift Accel-NGS 2S Plus Kit was compared to a traditional approach using an equivalent amount of input. A ChIP for the transcriptional repressor CTCF was performed on 50,000 cells, with half of the ChIP eluate used by a traditional library preparation and half used by the 2S Plus Kit. The resulting sequencing data illustrates the stronger and sharper peaks generated by the 2S Plus Kit. Additionally, data from the ENCODE Project is shown side-by-side to illustrate the simialrity of peak calls with the 2S Plus Kit, using 400-fold fewer cells as input.
CTCF ChIP-Seq from 50,000 Cells
CTCF ChIP-Seq was performed using 50,000 GM12878 lymphoblasts. The ChIP eluate was then split in half, with half of the material being made into a traditional library (purple) and the other half being made into a Swift Accel-NGS 2S Plus library (green), which reulted in stronger, sharper peaks. ENCODE ChIP-Seq data for CTCF is shown on the bottom track for comparison. Data generated by Active Motif.
Saturate Peak Calls from Less DNA
To explore the range of input levels required for comprehensive detection of precipitated regions, a titration curve was generated from 1 ng to 100 pg using a master ChIP sample prepared from mouse tissue. An antibody against the hematopoietic transcription factor PU.1 was used for the immunoprecipitation. DNA was quantified by Agilent Bioanalyzer and different amounts were used to prepare a series of NGS libraries that were sequenced on a HiSeq® 2500. When correcting for sequencing depth, we observed the same number of binding sites for samples prepared from 1 ng to 100 pg, demonstrating that 100 pg of starting material provides the same profile as 1 ng.
ChIP-Seq Sequencing Metrics with PU.1 Antibody
PU.1 ChIP sample was prepared from mouse bone marrow-derived macrophages. ChIP-Seq libraries prepared from different dilutions of the same ChIP DNA sample shared very similar sequencing metrics, including percentage of reads mapped, number of peaks identified, and mean peak score. The few peaks missing from the 250 pg and 100 pg samples relative to the 1 ng sample showed low mean peak scores, demonstrating proximity to background signal. Libraries were sequenced on a HiSeq 2500 using v4 chemistry and data were analyzed using Bowtie2, MACS 184.108.40.206, IntersecBed 2.17.0, and HOMER 3.1 tools.
Peak Comparison and Peak Quality for PU.1
Saturation curve for a range of ChIP DNA inputs. The saturation curve shared the same dynamics and the maximum number of peaks was reached at 40-50% of the aligned reads (75,000,000 reads) for all three inputs.
Peak Overlap Between Input Levels for PU.1