cell-free DNA methylation biomarker discovery in ALS

Mybaits probe capture design

Summer 2018

---

Project Goal

• Given three tissues of interest- brain, muscle, leukocytes- and CpG sites found using sparse PCA¹, design probes to capture cell free DNA that can illustrate tissue of origin between ALS and control cell-free DNA.

• List of CpG of interest contain ~100-1,000 CpGs determined to be differentially methylated given reference tissue profiling from ENCODE², Roadmap³, and other annotation projects.

Code Design

• Code designed to find all reads covering a CpG of interest in WGBS data (4 ALS and 4 CTRL merged)
• A SIZE parameter was used to search for the amount of base pairs around a CpG of interest to design probes for.
• Each CpG of interest was required to have at least 2 other CpGs to form coupled blocks of CpGs⁴

├── README.md
├── data
│   ├── muscle_cpgs_125.hg38.txt
├── output
│   ├── muscle_probes_125.txt
│   └── muscle_probes_125_patterns_all.txt
└── src
    ├── __init__.py
    ├── design
    │   ├── __init__.py
    │   ├── MethylRead.py
    │   └── cfDNA.py
    ├── run.py
    ├── run.sh
    └── utils
        ├── __init__.py
        └── io.py

Output

Probes file

• Has extension _probes_SIZE.txt where SIZE is the number of nucleotides around a CpG of interest assessed.
  • ex: muscle_probes_125.txt
• Contains 3 flavors of fasta entries per CpG of interest, 1) for reference sequence 2) for fully methylated sequence 3) read observed in the WGBS data
• Reference header:
  • parameters: search range around CpG | number of CpGs in region | tissue | reference
  • ex: >chr1:1093248-1093499|number_of_cpgs=5|tissue=muscle|reference
• Methylated header:
  • parameters: search range around CpG | number of CpGs in region | tissue | fully converted methylated
  • Fully converted means all C's in the sequence not part of a CpG dinucleotide are converted to a T
  • ex: >chr1:1093248-1093499|number_of_cpgs=5|tissue=muscle|fully_converted_methylated
• Read header:
  • parameters: Range read covers | number read in sequence
  • ex: >chr1:1093282-1093374|read_number=1
  • Alignment depicted to the reference strand. "." indicates a converted/non-methylated base (C-T mismatch for forward strand, G-A mistmatch for reverse strand)

Sample Entry:

>chr1:1093248-1093499|number_of_cpgs=5|tissue=muscle|reference
TGCTCCCTCTCTGGTTAAAGGGCATCCTGAGGGCCACATTAAGTCACAAAACATCATTTTGATTCAGGAACCAGAAGTCCAAGATTTCAATCAACACTTTCATCTGCTATTTAGTCAACTTCATGGAGATCTACTTTACATACAATAAACCACATCCGTGTAAAGTACACAAGCGGGTGAGTGTGACCACCCCCTTGAAGCTGCCACCACAGCCAGGACGGTGCCCGGTCCCACACAGCTGCCAGCACTCG
>chr1:1093248-1093499|number_of_cpgs=5|tissue=muscle|fully_converted_methylated
TGTTTTTTTTTTGGTTAAAGGGTATTTTGAGGGTTATATTAAGTTATAAAATATTATTTTGATTTAGGAATTAGAAGTTTAAGATTTTAATTAATATTTTTATTTGTTATTTAGTTAATTTTATGGAGATTTATTTTATATATAATAAATTATATTCGTGTAAAGTATATAAGCGGGTGAGTGTGATTATTTTTTTGAAGTTGTTATTATAGTTAGGACGGTGTTCGGTTTTATATAGTTGTTAGTATTCG
>chr1:1093282-1093374|read_number=1 |reverse_strand
TGCTCCCTCTCTGGTTAAAGGGCATCCTGAGGGCCACATTAAGTCACAAAACATCATTTTGATTCAGGAACCAGAAGTCCAAGATTTCAATCAACACTTTCATCTGCTATTTAGTCAACTTCATGGAGATCTACTTTACATACAATAAACCACATCCGTGTAAAGTACACAAGCGGGTGAGTGTGACCACCCCCTTGAAGCTGCCACCACAGCCAGGACGGTGCCCGGTCCCACACAGCTGCCAGCACTCG
                                  .|||||||.|||||||||||||||||.|||||..|||||.||.|||||.||||||||||||||||||||||.|||||||.||||||||||..                                                                                                                             
----------------------------------GACATTAAATCACAAAACATCATTTTAATTCAAAAACCAAAAATCCAAAATTTCAATCAACACTTTCATCTACTATTTAATCAACTTCATAA-----------------------------------------------------------------------------------------------------------------------------
>chr1:1093282-1093374|read_number=2|forward_strand
TGCTCCCTCTCTGGTTAAAGGGCATCCTGAGGGCCACATTAAGTCACAAAACATCATTTTGATTCAGGAACCAGAAGTCCAAGATTTCAATCAACACTTTCATCTGCTATTTAGTCAACTTCATGGAGATCTACTTTACATACAATAAACCACATCCGTGTAAAGTACACAAGCGGGTGAGTGTGACCACCCCCTTGAAGCTGCCACCACAGCCAGGACGGTGCCCGGTCCCACACAGCTGCCAGCACTCG
                                  .|.|||||||.|.||||.||.|||||||||.|||||..||||||..|||||||.|||.||.|.|||.||.||.||||||||.||.||.||||                                                                                                                             
----------------------------------TATATTAAGTTATAAAATATTATTTTGATTTAGGAATTAGAAGTTTAAGATTTTAATTAATATTTTTATTTGTTATTTAGTTAATTTTATGG-----------------------------------------------------------------------------------------------------------------------------

Patterns file

• Extension _patterns_all.txt
• Binary representation of the CpG patterning in a given region around a CpG of interest for all the reads covering that CpG
• This file illustrates the variability of the cfDNA reads covering a region
  • 0 = Unmethylated observation for read
  • 1 = Methylated observation for read
  • "-" = not covered by read
  • "." = read covered, but it is an incorrect base
• Header gives the genomic range covered, and % methylated for each CpG in locus. NA indicates no reads covered that CpG. " * " indicates CpG of interest
  • Ex: > chr1:7224041-7224292(cpg*1*: 0.781, cpg2: 0.962, cpg3: NA)

Sample entry:

> chr1:110347251-110347502(cpg1: 1.0, cpg*2*: 0.891, cpg3: 0.985, cpg4: 0.969, cpg5: 0.99, cpg6: 0.969, cpg7: 1.0)
111----
111----
111----
111----
111----
111----
111----
111----
1111---
1010---
1111---
1011---
1011---
1111---
1111---
1111---
--10111
--11111
--.1111
--.1111
--11111
--11111

References

1 Rahmani et al. Nat Methods 2017 "Sparse PCA Corrects for Cell-Type Heterogeneity in Epigenome-Wide Association Studies ". ↩

2 ENCODE Consortium, 2012, ENCODE encyclopedia, Version 4: Genomic Annotations. ↩

3 NIH Roadmap Epigenomics Mapping Consortium, 2015, NIH Roadmap Epigenomics Project Data Listings . ↩

4 Guo et al, Nat Gen, 2017, Identification of methylation haplotype blocks aids in deconvolution of heterogeneous tissue samples and tumor tissue-of-origin mapping from plasma DNA . ↩