To meet the need for a reference control in cell-free methylated DNA immunoprecipitation-sequencing (cfMeDIP-seq)1,2 experiments, we designed spike-in controls and ligated unique molecular indexes (UMI) to adjust for PCR bias, and immunoprecipitation bias caused by the fragment length, G+C content, and CpG density of the DNA fragments that are immunoprecipitated3. This enables absolute quantification of methylated DNA in picomoles, while retaining epigenomic information that allows for sensitive, tissue-specific detection as well as comparable results between different experiments. We designed DNA fragments with 2x3x3x3=54 combinations of methylation status (methylated and unmethylated), fragment length in basepair (bp) (80 bp, 160 bp, 320 bp), G+C content (35%, 50%, 65%), and fraction of CpGs within a fragment (1 CpG/ 80 bp, 1 CpG/ 40 bp, 1 CpG/ 20 bp). Spiky was developed for analyzing DNA methylation of cell-free DNA obtained from cfMeDIP-seq method using reference ‘spike-in’ controls. This package will:
Install and load the spiky package from Bioconductor.
sequence (DNAStringSet) methylated (boolean) CpGs (integer) fmol (numeric) molmass (numeric) GCfrac (numeric) OECpG (numeric) conc (numeric) *frag_grp (character)
If you are using the same synthetic spike-in sequences as described in the manuscript, you may load the spike-in sequence database using:
If you are using custom spike-ins, you can create your own spike-in sequence database using the process_spikes() function, which accepts as input a Fasta file, GRanges, or dataframe, and a vector of booleans (0 or 1) describing whether each spike-in sequence is methylated. The input Fasta file can also be generated from BAM header info as follows:
sb <- system.file("extdata", "example.spike.bam", package="spiky", mustWork=TRUE)
outFasta <- paste(system.file("extdata", package="spiky", mustWork=TRUE),"/spike_contigs.fa",sep="")
show(generate_spike_fasta(sb, spike=spike,fa=outFasta))
#> Wrote /tmp/RtmpimFLt1/Rbuild1ff9729687d2/spiky/inst/extdata/spike_contigs.fa
#> DNAStringSet object of length 52:
#> width seq names
#> [1] 160 CTTTACTACTGAATGTAAGCTCT...TTTACTATAACCGATTACACAT 160b_2C_35G-2_meth
#> [2] 160 GTAACATGGTTACCACTGGGACC...ACTAGCCGTGTCCCAACCTCAT 160b_2C_50G-2_meth
#> [3] 160 GACTCCTCCCTAGGCCCCCATGG...CATGCAGGCCCCCCGCTCCATC 160b_2C_65G-2_meth
#> [4] 160 GTATAATCATAACAAAGGCCTAA...AGCGACTAAGATCTCAGAATTA 160b_4C_35G-2_mod...
#> [5] 160 AGCCTTGGACGTGAGTCTCTGTT...ACAATTGTCAGGGCCCTCCAGT 160b_4C_50G-1_meth
#> ... ... ...
#> [48] 80 ACAACACCCTCCACCCAATACTT...AAGTCAGTCAAATGCCTGTAAC 80b_2C_50G-1
#> [49] 80 CACCTTGAGACCTCCAGAGGGGG...GTGCGCAGGGGGGAAGGGGGGC 80b_2C_65G-1
#> [50] 80 AAGGCATTACTTATCTAATCAAT...TTTGTACTCGTAGACGAAATTG 80b_4C_35G-1
#> [51] 80 AGTCATCAGCATATTGTCAGTAC...TACTCCTAGTGGGCTGCGTGGT 80b_4C_50G-1
#> [52] 80 TTGGGAGGCTCTGGACTGGGGCA...CGTCCCCCCCCATCCTCTCCGC 80b_4C_65G-1
The spike-in database can then be created with this input Fasta.
spikes <- system.file("extdata", "spikes.fa", package="spiky", mustWork=TRUE)
spikemeth <- spike$methylated
process_spikes(spikes, spikemeth)
#> DataFrame with 52 rows and 5 columns
#> sequence methylated CpGs GCfrac OECpG
#> <DNAStringSet> <numeric> <integer> <numeric> <numeric>
#> 80b_1C_35G-1 TAGGATATAG...AAATTATGCT 0 1 0.35 0.29
#> 80b_1C_35G-2 TGTCTAAATT...AAGAACATAT 1 1 0.35 0.29
#> 80b_2C_35G-1 TCTAATACTC...AATCCATAAG 0 2 0.35 0.57
#> 80b_2C_35G-2 CTCAAATATA...CAATAACACT 1 2 0.35 0.57
#> 80b_4C_35G-1 AAGGCATTAC...GACGAAATTG 0 4 0.35 1.14
#> ... ... ... ... ... ...
#> 320b_4C_65G-2 GCAATTGATG...AAGAAGCTAA 0 4 0.35 0.29
#> 320b_8C_65G-1 CCCATGCATC...TCTTACCAGT 1 8 0.50 0.40
#> 320b_8C_65G-2 GAACTTCCAA...ATGCTATGCA 0 8 0.50 0.40
#> 320b_16C_65G-1 TTCGGCACTT...GCTTAAGAAA 1 16 0.50 0.80
#> 320b_16C_65G-2 TTGGGCCGCC...CTGAGATTGT 0 16 0.50 0.80
Spiky supports input files in either BAM or BEDPE format.
BAM file in standard format (For full details about the BAM format, see https://samtools.github.io/hts-specs/SAMv1.pdf). The BAM must also have an accompanying index file, which can be created using samtools index ${filename.bam}. (http://www.htslib.org/doc/samtools-index.html)
genomic_bam_path <- system.file("extdata", "example_chr21.bam", package="spiky", mustWork=TRUE)
genomic_coverage <- scan_genomic_contigs(genomic_bam_path,spike=spike)
#> Warning in find_spike_contigs(si, spike = spike): Cannot resolve contig name mismatches with spike database names. Your file may not have any spikes, or the spike database might not contain them.
#> Tiling 300bp bins across the genome...Done.
#> Binning genomic coverage...
#> Binning genomic coverage...Done.
#> Done.
spike_bam_path <- system.file("extdata", "example.spike.bam", package="spiky", mustWork=TRUE)
spikes_coverage <- scan_spike_contigs(spike_bam_path,spike=spike)
#> Summarizing spike-in counts...Done.
BEDPE file in standard format. For full details about the BEDPE format, see Bedtools documentation (https://bedtools.readthedocs.io/en/latest/content/general-usage.html#bedpe-format). In short, for a pair of ranges 1 and 2, we have fields chrom1, start1, end1, chrom2, start2, end2, and (optionally) name, score, strand1, strand2, plus any other user defined fields that may be included (these are not yet supported by Spiky). For example, two valid BEDPE lines are:
chr1 100 200 chr5 5000 5100 bedpe_example1 30
chr9 900 5000 chr9 3000 3800 bedpe_example2 99 + -
The BEDPE must also have an accompanying index file, which can be created using Bedtools, as in the example shown below, where ${file} represents the name of a BEDPE file.
bedtools sort -i ${file} > sorted_${file} bgzip sorted_${file} tabix sorted_${file}.gz
genomic_bedpe_path <- system.file("extdata", "example_chr21_bedpe.bed.gz", package="spiky", mustWork=TRUE)
genomic_coverage <- scan_genomic_bedpe(genomic_bedpe_path,genome="hg38")
#> Tiling 300bp bins across the genome...Done.
#> Binning genomic coverage...
#> Binning genomic coverage...Done.
#> Done.
spike_bedpe_path <- system.file("extdata", "example_spike_bedpe.bed.gz", package="spiky", mustWork=TRUE)
spikes_coverage <- scan_spike_bedpe(spike_bedpe_path,spike=spike)
#> Warning in .toPairs(res): NAs introduced by coercion
#> Summarizing spike-in counts...Done.
For each combination of parameters, we designed two distinct spike-in sequences. One to be methylated and one to be unmethylated. The allows us to assess non-specific binding of the monoclonal antibody on a sample-by-sample basis. To calculate methylation specificity we take the number of methylated reads divided by the total number of reads. It is our recommendation that if methylation specificity is <0.98, then the sample should be flagged or removed from analysis as the cfMeDIP performed inadequately.
This calculation is done by the ‘methylation_specificity’ function.
For each batch of samples, the coefficients used in the Gaussian generalized linear model will differ. The ‘model_glm_pmol’ will calculate these coefficients and output the model to be used to calculate molar amount (picomoles) on the user’s DNA sequences of interest. We assume that all DNA sequences of interest are methylated after undergoing cfMeDIP-seq. As such, we build the Gaussian generalized linear model on only the methylated spike-in control fragments. A generated Bland-Altman plot will visualize how well the model performs.
## Build the Gaussian generalized linear model on the spike-in control data
gaussian_glm <- model_glm_pmol(spikes_coverage,spike=spike)
#> Converting x (a coverage result?) to a data.frame...
summary(gaussian_glm)
#>
#> Call:
#> glm(formula = conc ~ read_count + fraglen + GC + CpG_3, data = x)
#>
#> Coefficients:
#> Estimate Std. Error t value Pr(>|t|)
#> (Intercept) 3.914e-03 4.919e-04 7.958 8.96e-08 ***
#> read_count -1.946e-08 9.913e-09 -1.964 0.063 .
#> fraglen -1.149e-05 1.295e-06 -8.873 1.50e-08 ***
#> GC 3.702e-06 7.407e-06 0.500 0.622
#> CpG_3 1.719e-04 2.894e-04 0.594 0.559
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> (Dispersion parameter for gaussian family taken to be 1.733019e-07)
#>
#> Null deviance: 3.3927e-05 on 25 degrees of freedom
#> Residual deviance: 3.6393e-06 on 21 degrees of freedom
#> AIC: -324.54
#>
#> Number of Fisher Scoring iterations: 2
For the samples in which the Gaussian generalized linear model was built, we will calculate the molar amount (picomoles) for each DNA sequence of interest.
# Predict pmol concentration
# To select a genome other than hg38, use BSgenome::available.packages() to find valid BSgenome name
#library("BSgenome.Hsapiens.UCSC.hg38")
sample_data_pmol <- predict_pmol(gaussian_glm, genomic_coverage,bsgenome="BSgenome.Hsapiens.UCSC.hg38",ret="df")
#> Adjusting for bin-level biases...
#> Attempting to load BSgenome.Hsapiens.UCSC.hg38...
#> OK.
#> Done.
#> Starting pmol prediction...
#> Done.
head(sample_data_pmol,n=1)
#> chrom range.start range.end
#> chr21:30201601-30201900 chr21 30201601 30201900
#> sequence
#> chr21:30201601-30201900 AGGAAGGGGTTCAGTTTCAGTTTTCTGCATATGGCTAGCCAGTTTTCTCAACACCTTTTATTAAATAGGGAATCATTTCCCCATTGCTTGTTTTTGTCAGGTTTGTCAAAGATGAGATGGTTGTAGACGTGCGGTGTTATTTCTGAGGCCTCTGTTCTGTTCCATTGGTCTATATATCTGTTTTGGTATCAGTATCATGCTGTTTTTGTTACTGTAGCCTTGTAGTATAGTTTGAAGTCAGGTAGTGTGATGCCTCCAGCTTTGTTCTTTTTGTTTAGAATTGTCTTGGCTATACAGGCT
#> read_count fraglen GC CpG_3 pred_conc
#> chr21:30201601-30201900 1.8 300 0.3933333 1.259921 0.0006848288
For our analyses, we binned the genome into 300 bp non-overlapping windows. We then look overlap between fragments in our data with each of the 300 bp genomic windows. We adjust the molar amount (picomoles) by a multiplier. This multiplier is the proportion of overlap between our fragment and the 300 bp window. This is done for every fragment in our sample.
sample_binned_data <- bin_pmol(sample_data_pmol)
head(sample_binned_data,n=1)
#> chrom range.start range.end
#> chr21:30201601-30201900 chr21 30201601 30201900
#> sequence
#> chr21:30201601-30201900 AGGAAGGGGTTCAGTTTCAGTTTTCTGCATATGGCTAGCCAGTTTTCTCAACACCTTTTATTAAATAGGGAATCATTTCCCCATTGCTTGTTTTTGTCAGGTTTGTCAAAGATGAGATGGTTGTAGACGTGCGGTGTTATTTCTGAGGCCTCTGTTCTGTTCCATTGGTCTATATATCTGTTTTGGTATCAGTATCATGCTGTTTTTGTTACTGTAGCCTTGTAGTATAGTTTGAAGTCAGGTAGTGTGATGCCTCCAGCTTTGTTCTTTTTGTTTAGAATTGTCTTGGCTATACAGGCT
#> read_count fraglen GC CpG_3 pred_conc
#> chr21:30201601-30201900 1.8 300 0.3933333 1.259921 0.0006848288
#> adjusted_pred_con
#> chr21:30201601-30201900 0.001232692
sessionInfo()
#> R version 4.4.2 (2024-10-31)
#> Platform: x86_64-pc-linux-gnu
#> Running under: Ubuntu 24.04.1 LTS
#>
#> Matrix products: default
#> BLAS: /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
#> LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.26.so; LAPACK version 3.12.0
#>
#> locale:
#> [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
#> [3] LC_TIME=en_US.UTF-8 LC_COLLATE=C
#> [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
#> [7] LC_PAPER=en_US.UTF-8 LC_NAME=C
#> [9] LC_ADDRESS=C LC_TELEPHONE=C
#> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
#>
#> time zone: Etc/UTC
#> tzcode source: system (glibc)
#>
#> attached base packages:
#> [1] stats4 stats graphics grDevices utils datasets methods
#> [8] base
#>
#> other attached packages:
#> [1] BSgenome.Hsapiens.UCSC.hg38_1.4.5 spiky_1.13.0
#> [3] Rsamtools_2.23.0 Biostrings_2.75.1
#> [5] XVector_0.47.0 GenomicRanges_1.59.1
#> [7] GenomeInfoDb_1.43.1 IRanges_2.41.1
#> [9] S4Vectors_0.45.2 BiocGenerics_0.53.3
#> [11] generics_0.1.3 devtools_2.4.5
#> [13] usethis_3.0.0 rmarkdown_2.29
#>
#> loaded via a namespace (and not attached):
#> [1] bitops_1.0-9 remotes_2.5.0
#> [3] rlang_1.1.4 magrittr_2.0.3
#> [5] distributions3_0.2.2 matrixStats_1.4.1
#> [7] compiler_4.4.2 mgcv_1.9-1
#> [9] vctrs_0.6.5 profvis_0.4.0
#> [11] pkgconfig_2.0.3 crayon_1.5.3
#> [13] fastmap_1.2.0 ellipsis_0.3.2
#> [15] utf8_1.2.4 promises_1.3.0
#> [17] sessioninfo_1.2.2 UCSC.utils_1.3.0
#> [19] purrr_1.0.2 xfun_0.49
#> [21] zlibbioc_1.52.0 cachem_1.1.0
#> [23] jsonlite_1.8.9 later_1.3.2
#> [25] DelayedArray_0.33.2 BiocParallel_1.41.0
#> [27] parallel_4.4.2 R6_2.5.1
#> [29] bslib_0.8.0 rtracklayer_1.67.0
#> [31] pkgload_1.4.0 jquerylib_0.1.4
#> [33] Rcpp_1.0.13-1 SummarizedExperiment_1.37.0
#> [35] knitr_1.49 BlandAltmanLeh_0.3.1
#> [37] httpuv_1.6.15 Matrix_1.7-1
#> [39] splines_4.4.2 rstudioapi_0.17.1
#> [41] abind_1.4-8 yaml_2.3.10
#> [43] codetools_0.2-20 miniUI_0.1.1.1
#> [45] curl_6.0.1 pkgbuild_1.4.5
#> [47] lattice_0.22-6 tibble_3.2.1
#> [49] Biobase_2.67.0 shiny_1.9.1
#> [51] withr_3.0.2 coda_0.19-4.1
#> [53] evaluate_1.0.1 desc_1.4.3
#> [55] survival_3.7-0 urlchecker_1.0.1
#> [57] pillar_1.9.0 MatrixGenerics_1.19.0
#> [59] rprojroot_2.0.4 sp_2.1-4
#> [61] RCurl_1.98-1.16 ggplot2_3.5.1
#> [63] munsell_0.5.1 scales_1.3.0
#> [65] xtable_1.8-4 glue_1.8.0
#> [67] maketools_1.3.1 tools_4.4.2
#> [69] MBA_0.1-2 BiocIO_1.17.0
#> [71] sys_3.4.3 BSgenome_1.75.0
#> [73] GenomicAlignments_1.43.0 buildtools_1.0.0
#> [75] fs_1.6.5 mvtnorm_1.3-2
#> [77] XML_3.99-0.17 grid_4.4.2
#> [79] colorspace_2.1-1 nlme_3.1-166
#> [81] GenomeInfoDbData_1.2.13 restfulr_0.0.15
#> [83] Formula_1.2-5 cli_3.6.3
#> [85] fansi_1.0.6 S4Arrays_1.7.1
#> [87] bamlss_1.2-5 gtable_0.3.6
#> [89] sass_0.4.9 digest_0.6.37
#> [91] SparseArray_1.7.2 rjson_0.2.23
#> [93] htmlwidgets_1.6.4 memoise_2.0.1
#> [95] htmltools_0.5.8.1 lifecycle_1.0.4
#> [97] httr_1.4.7 mime_0.12