SPRITE

Protocol

Split-Pool Recognition of Interactions by Tag

Extension (SPRITE) for DNA: Experimental
Protocols

Sofia Quinodoz
squinodo@lncrna.caltech.edu
Mitchell Guttman Laboratory
Last edited: January 21, 2018

This document describes the experimental procedures for our SPRITE (Split-Pool
Recognition of Interactions by Tag Extension) method to be used for mapping
genome-wide higher order interactions between DNA molecules.

1
 SPRITE Protocol

Table of Contents:

1 Materials
1.1 Solutions
1.2 Equipment
1.3 Additional Materials and Reagents

2 Cell Culture

3 Adaptor and Barcode Design
3.1 DNA Phosphate Modified (DPM) Adaptor
3.2 Odd and Even Tags
3.3 Terminal Tag
3.4 Final Library Amplification
3.5 DPM primers for QC of DPM ligation
3.6 Adaptor annealing program

4 Sample Preparation
4.1 Formaldehyde-DSG Crosslinking
4.2 Cell Lysis
4.3 DNA Fragmentation

5 SPRITE Pre Split-and-Pool and QC Pt. 1
5.1 NHS Coupling
5.2 Phosphorylation and End Repair
5.3 DPM Adaptor Ligation
5.4 QC: Check to Determine Ligation Efficiency of the DPM Adaptor

6 SPRITE and Library Preparation Pt. 2
6.1 SPRITE
6.2 Library Preparation Pt. 2
6.3 Estimating Sequencing Depth

7 Sequencing and Data Analysis
7.1 Tag identification
7.2 Alignment
7.3 Filtration
7.4 Subsequence post-processing
7.5 Quality Controls of Successful SPRITE Libraries

2
 SPRITE Protocol

1 Materials

1.1 Solutions

DSG Crosslinking Cell Lysis Buffer C Coupling Buffer
Solution 10mM Tris pH 8 1X PBS
1X PBS 1.5mM EDTA 0.1% SDS
2mM DSG in DMSO 1.5mM EGTA
100mM NaCl RLT++ Buffer
Scraping Buffer 0.1% DOC 1X Buffer RLT supplied
1 x PBS pH 7.5 0.5% NLS by Qiagen
0.5% BSA 10mM Tris pH 7.5
Store at 4 °C 10x DNase Buffer 1mM EDTA
200mM Hepes pH 7.4 1mM EGTA
Cell Lysis Buffer A 1M NaCl 0.2% NLS
50mM Hepes pH 7.4 0.5% NP-40 0.1% Triton-X
1mM EDTA 5mM CaCl2 0.1% NP-40
1mM EGTA 25mM MnCl2
140mM NaCl M2 Wash Buffer
0.25% Triton-X 25x DNase Stop 20mM Tris pH 7.5
0.5% NP-40 Solution 50mM NaCl
10% Glycerol 250mM EDTA 0.2% Triton-X
125mM EGTA 0.2% NP-40
Cell Lysis Buffer B 0.2% DOC
10mM Tris pH 8 MyRNK Buffer
1.5mM EDTA 20mM Tris pH 7.5 PBLSD+ Wash Buffer
1.5mM EGTA 100mM NaCl 1X PBS
200mM NaCl 10mM EDTA 5mM EDTA
10mM EGTA 5mM EGTA
10x Annealing Buffer 0.5% Triton-X 5mM DTT (add fresh)
100mM Tris-HCl pH 7.5 0.2% SDS 0.2% Triton-X
2M LiCl 0.2% NP-40
2mM EDgTA 0.2% DOC



Note 1: RLT++ Buffer contains guanidine thiocyanate which when mixed with
bleach produces hydrogen cyanide gas and hydrogen chloride gas. Be careful to
ensure that all liquid RLT++ Buffer waste is disposed of in its own waste container.
Solids that have touched RLT++ Buffer such as tips and reservoirs should also be
discarded in a separate solid RLT++ Buffer container.

Note 2: DTT has a short half-life at pH 7.4 at 20C. It is important to keep PBLSD+
Buffer on ice during the procedure and frozen at -20C if not in use.

3
 SPRITE Protocol

4
 SPRITE Protocol

1.2 Equipment

Microcentrifuge
Plate Centrifuge
Sonication instrument and chiller
Gel Electrophoresis Equipment
Qubit Fluorometer
Eppendorf Thermomixer
Eppendorf SmartBlock 1.5mL thermoblock
Eppendorf SmartBloack PCR 96 thermoblock
Magnetic rack for 1.5mL tubes (e.g. Invitrogen DynaMag-2)
Magnetic rack for 15mL conical tubes
Magnetic rack for 96 well plate
PCR machine
Agilent Bioanalyzer

5
 SPRITE Protocol

1.3 Additional Materials and Reagents

Barcode and adaptors
Low-retention pipette tips
Low-bind 96-well plate
Reservoirs
PCR Strip tubes
Protein Lo-bind 1.5ml eppendorf tubes
15mL Falcon conicals
50mL Falcon conicals
TVP
PBS Sterile
Disuccinimidyl glutarate (DSG), 50mg bottle from Pierce, brought to 0.5M with the
addition of DMSO
16% Formaldehyde Solution Ampules from Pierce
2.5M Glycine
Protease Cocktail Inhibitor
TURBO DNase from ThermoFisher Scientific
Proteinase K from New England Biolabs
DNA Clean and Concentrator-5 Kit with Capped Columns from Zymo Research
Gel Electrophoresis System
Pierce NHS-Activated Magnetic Beads
100mM ATP from New England Biolabs
T4 Polynucleotide Kinase from New England Biolabs
Polynucleotide Kinase Reaction Buffer from New England Biolabs
NEBNext End Repair Enzyme Mix from New England Biolabs
NEBNext End Repair Reaction Buffer from New England Biolabs
Klenow Fragment (3’ to 5’ exo-) From New England Biolabs
NEBNext dA-Tailing Reaction Buffer From New England Biolabs
Instant Sticky-end Ligase Master Mix from New England Biolabs
Q5 Hot Start High-Fidelity 2X Master Mix from New England Biolabs
Agencourt AMPure XP Magnetic Beads from Beckman Coulter
5’ Deadenylase from New England Biolabs
Qubit dsDNA HS Assay Kit
Agilent High Sensitivity DNA Kit

6
 SPRITE Protocol

2 Cell Culture

Materials and Methods

Mouse ES cell culture and Xist induction
All mouse ES cell lines were cultured in serum-free 2i/LIF medium as previously
described (7, 11, 13). Female ES cells (F1 2-1 line, generously provided by K. Plath)
are an F1 hybrid wild-type mouse ES cell line derived from a 129 × CAST
(castaneous) cross. Maintenance of 2 X chromosomes in this line was monitored by
X chromosome paint imaging, restriction length polymorphism analysis, as well as
Sanger sequencing of SNPs on the X chromosome. The pSM33 ES cell line (kindly
provided by K. Plath) is a male ES cell line, derived from the V6.5 ES cell line,
expressing the lncRNA Xist from the endogenous locus under the transcriptional
control of a tet-inducible promoter and the Tet transactivator (M2rtTA) from the
Rosa26 locus. To induce Xist, doxycycline (Sigma, D9891) was added to cultures at a
final concentration of 2ug/ml for 6-24hrs.

Human lymphoblast cell culture
GM12878 cells (Coriell Cell Repositories), a human lymphoblastoid cell line, was
cultured in RPMI 1640 (Gibco, Life Technologies), 2mM L-glutamine, 15% fetal
bovine serum, and 1x penicillin-streptomycin and maintained at 37°C under 5%
CO2. Cells were seeded every 3-4 days at 200,000 cells/ml in T25 flasks and
passaged or harvested before reaching 1,000,000 cells/ml.

References
1. R. Galupa, E. Heard, X-chromosome inactivation: new insights into cis and trans regulation. Curr. Opin. Genet. Dev.
31, 57–66 (2015). 

2. E. Splinter et al., The inactive X chromosome adopts a unique three-dimensional conformation that is dependent on
Xist RNA. Genes Dev. 25, 1371–1383 (2011). 

3. S. S. Rao et al., A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell.
159, 1665–1680 (2014). 

4. A. Rego, P. B. Sinclair, W. Tao, I. Kireev, A. S. Belmont, The facultative heterochromatin of the inactive X chromosome
has a distinctive condensed ultrastructure. J Cell Sci. 121, 1119–1127 (2008). 

5. A. Wutz, Gene silencing in X-chromosome inactivation: advances in understanding facultative heterochromatin
formation. Nat. Rev. Genet. 12, 542–553 (2011). 

6. C. M. Clemson, L. L. Hall, M. Byron, J. McNeil, J. B. Lawrence, The X chromosome is organized into a gene-rich outer
rim and an internal core containing silenced nongenic sequences. Proc. Natl. Acad. Sci. U. S. A. 103, 7688–7693
(2006). 

7. J. M. Engreitz et al., The Xist lncRNA exploits three-dimensional genome architecture to spread across the X
chromosome. Science (80-89). 341, 1237973 (2013). 

8. M. D. Simon et al., High-resolution Xist binding maps reveal two-step spreading during X-chromosome inactivation.
Nature. 504, 465–469 (2013).
9. J. Chaumeil, P. Le Baccon, A. Wutz, E. Heard, A novel role for Xist RNA in the formation of a repressive nuclear
compartment into which genes are recruited when silenced. Genes Dev. 20, 2223–2237 (2006). 

10. A. Wutz, T. P. Rasmussen, R. Jaenisch, Chromosomal silencing and localization are mediated by different domains of
Xist RNA. Nat. Genet. 30, 167–174 (2002). 

11. C. A. McHugh et al., The Xist lncRNA interacts directly with SHARP to silence transcription through HDAC3. Nature.
521, 232–236 (2015). 

12. C. Chu et al., Systematic discovery of Xist RNA binding proteins. Cell. 161, 404–416 (2015).
13. C. Chen et al., Xist recruits the X chromosome to the nuclear lamina to enable chromosome-wide silencing. Science.
354, 468-472 (2016).

7
 SPRITE Protocol

3 Adaptor and Barcode Design

The above figure demonstrates the adaptor and tag scheme that is central to the
SPRITE process. SPRITE uses a split-and-pool strategy to uniquely barcode all
molecules within a crosslinked complex by repeatedly splitting all complexes into a
96-well plate, ligating a specific tag sequence within each well, followed by pooling
of these complexes such that the final product contains a series of tags ligated to
each molecule, which we refer to as a barcode.


3.1 DNA Phosphate Modified (DPM) Adaptor

5'Phos AAACACCCAAGATCGGAAGAGCGTCGTGTA 3’ Spcr
||||||||||||||||||||
3' TTTTGTGGGTTCTAGCCTTCTGTACTGTTCAGT 5’Phos


The above dsDNA molecule is an example of one of the 96 DPM adaptors used
during our process. The 5’ end of the molecule has a modified phosphate group that
allows for the ligation between DPM and the target DNA molecules as well as the
subsequent tag. The highlighted regions on DPM have the following functions:

1. The yellow T overhang is a sticky-end that ligates to our target DNA
molecules, which are given a 5’ A overhang following end repair.
2. The pink region is the 9-nucleotide sequence unique to each of the 96 DPM
adaptors. These unique sequences help to identify post-sequencing DNA
molecules that are in a complex.
3. The green sequence is a sticky end that ligates to the first tag.
4. The grey sequence is complementary to the First Primer used for library
amplification. Part of the grey sequence makes up a 3’ spacer to prevent the
top strand of the Odd tag from ligating, and only the bottom
5’phosphorylated sticky end of the Odd tag will ligate to the green tag. Its
purpose is discussed in section 3.4.

8
 SPRITE Protocol

3.2 Odd and Even Tags

Odd and Even tags are so named because the Odd tag is ligated 1st, 3rd, 5th, etc…
during the SPRITE process and the Even tag is ligated 2nd, 4th, 6th, etc… during
SPRITE for however many rounds of tagging and pooling are completed. It is not
necessary to ligate only an even number of tags or only an odd number of tags so
long as there are two sets of Terminal tags; one that can ligate to Odd tags and one
that can ligate to Even tags.

5'Phos CAAGTCAAGCTAGATTCCACGAAGAGTTGTCACGTCAGCCGCAGTATC 3’
|||||||||||||||||||||||||||||||||||||||||
3' TCGATCTAAGGTGCTTCTCAACAGTGCAGTCGGCGTCATAGGTTCAGT 5’Phos


The above dsDNA molecule is an Odd tag and an Even tag ligated together. The
following points are important to note:
1. The 5’ overhang on the top strand ligates either to the DPM adaptor (green
sequence in section 3.1) or the 5’ overhang on the bottom strand of the Even
tag.
2. Both the Odd tags and Even tags have modified 5’ phosphate groups to allow
for tag elongation.
3. The bolded regions of complementarity on each tag are the sequences
unique to each of the 96 tags (192 total, accounting to Odd tags and Even
tags).

3.3 Terminal Tag

The terminal tag below ligates to Odd tags, though a terminal tag has also been
made to ligate to Even tags. The key feature of the terminal tag is that there is no
modified 5’ phosphate on the bottom strand.

5' Phos AGTTGTCACCATAATAAGATCGGAAGA 3’

||||||||||||||||||||
3' TGGTATTATTCTAGCCTTCTCGTGTGCAGAC 5’

1. The grey sequence is complementary to the Second Primer used for library
amplification.
2. However since DNA cannot be synthesized in a 3’ to 5’ direction, the Second
Primer anneals to a daughter strand synthesized from the First Primer, as
explained in section 3.4. The top strand is not primed because there is a
break in the sequence generated by the 3’spacer on the DPM molecule and
therefore priming the top strand of the terminal tag would terminate at the
barcodes and would not PCR through to the gDNA sequence ligated to the
barcodes.
3. The bolded sequence on the Terminal tag is unique to each of the 96 tags.

9
 SPRITE Protocol

3.4 Final Library Amplification

The DPM adaptor is designed with a 3’ spacer to aid in final library amplification. If
the 3’ spacer is absent, each strand will form a hairpin loop during the initial
denaturation due to reverse complementarity of the sequences on either side of the
target DNA molecule. Instead, the 3’ spacer allows the barcodes to only ligate to the
5’end of each single-stranded DNA sequence, and not the 3’end, preventing these
hairpin from forming.

2P_universal (F primer)
5’ AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT 3’

2P_barcoded_85 (R primer)
5’ CAAGCAGAAGACGGCATACGAGATGCCTAGCCGTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT 3’

Due to reverse complementarity of the sequences, only one primer amplifies the
tagged DNA in the first PCR cycle. This First Primer anneals to a sequence in the
DPM adaptor and extends, synthesizing two daughter strands with reverse
sequences. This first primer serves as the Read1 primer during Illumina
sequencing. To synthesize the complement, the Second Primer anneals to the
daughter strand extended from the First Primer in the second PCR cycle.

The 2P_barcoded primer contains an 8 nucleotide barcode within the primer. This
barcode is read from the illumina sequencer during the indexing priming step. This
barcode effectively serves as an additional round of tag addition during SPRITE.
Dilution of the sample into multiple wells is performed at the final step of SPRITE
prior to proteinase K elution from NHS beads. Each dilution of the sample prior to
proteinase K elution isolates a subset of the tagged complexes into different wells.
Each dilution of complexes are amplified with a different 2P_barcoded primer.

Both the First and Second primers are around 30 nucleotides each. Yet the
sequences they anneal to initially are ~20 nucleotides. For this reason, we set two
different annealing temperatures during the final library PCR. The first annealing
temperature is for the first four cycles until enough copies are made with fully
extended primer regions. After these four cycles, the annealing temperature is
raised for a remaining five cycles.

The 2P_universal primer and 2P_barcoded serve as the Read 1 and Read 2 primers
for illumina sequencing, respectively. Read 1 sequences the DNA molecule and the
DPM adaptor. Read 2 sequences the multiple tags, ie. unique barcode, ligated to the
DNA molecules.


3.5 DPM primers for Quality-Control of DPM ligation

These primers are used to ensure that the DPM adaptor has been successfully
ligated to DNA of the lysate. If no libraries are obtained at this step after 14-16

10
 SPRITE Protocol

cycles of PCR, we strongly recommend to not proceed as subsequent ligation of tags

and amplification of tagged DNA during the SPRITE protocol will be unsuccessful.

DPMQCprimerF 5’ TACACGACGCTCTTCCGATCT 3’
DPMQCprimerR 5’ TGACTTGTCATGTCTTCCGATCT 3’

The Forward and Reverse primers amplify the top strand and bottom strand of the
DPM adaptor, respectively (section 3.1).


3.5 Adaptor annealing program
The following adaptors are annealed to make the tags double-stranded adaptors for
dsDNA adaptor ligation:
1. DPM adaptors
2. Odd adaptors
3. Even adaptors
4. Terminal Tag adaptors

Mix the top and bottom strands of each adaptor into a PCR tube or 96-well plate
with 10x Annealing Buffer:
Reagents Volume
10x Annealing Buffer 10ul
Top Adaptor (200μM) 45ul
Bottom Adaptor (200μM) 45ul
Total 100ul

Incubate with the following conditions in a thermocycler for adapter annealing to
denature any secondary structure within the top and bottom strands of each
adaptor, then slowly cool to anneal each strand:
Temperature (°C) Time (min) Ramp (°C/s) Cycle
Denaturation 95 02:00
Annealing 85 00:10 -1 60
Hold 25 Infinite

11
 SPRITE Protocol

4 Sample Preparation

Goal: Crosslink cells to fix in vivo RNA-DNA-Protein complexes with disuccinimidyl
glutarate (DSG) and formaldehyde crosslinkers. Lyse cells and fragment DNA to
appropriate sizes via sonication and DNase.

Optimization of lysis conditions (amount of sonication, amount/timing of DNase) is
a critical step in establishing the protocol for the first time. The length of sonication
might vary from 30sec to several minutes and DNase treatment might vary from 10
to 20 minutes, depending on cell number, ploidy, crosslinking strength, and the
desired DNA fragment size. To optimize DNase timing and conditions, remove 5 µL
lysate aliquots every 2-4 minutes, quench with EDTA and EGTA on ice, and assay
DNA sizes for each time point as described in the protocol. If an appropriate
combination of solubilization and DNA fragment sizes cannot be obtained by
varying the amount of sonication or DNase, then reducing the strength of the
crosslinking may be necessary.(1)

4.1 Formaldehyde-DSG Crosslinking

1. Grow adherent cells on 15-cm plates. Before crosslinking, count one plate.
This protocol details crosslinking multiple plates of cells in one suspension,
but it is important to maintain consistency in lysate batches. We typically
store cells in 10M pellets.

2. An hour before starting, warm TVP and wash solution at 37C. Chill one bottle
of PBS at 4C, keep one at room temperature.

3. Lift cells from plate and wash: Remove media from plates. Add 5mL TVP to
each 15cm plate and rock gently for 3-4 minutes. Afterwards, add 25mL
wash solution to each plate. Vigorously suspend cells in the wash solution
and transfer from plate to a 50mL conical tube. Rinse the plate with extra
wash solution and add to the 50mL conical. Pellet in a centrifuge for 3
minutes at 3300 X G at room temperature. Wash cells by resuspending in
4mL room temperature 1X PBS per 10M cells and transfer to a 15mL conical.
Pellet again.

4. Resuspend cells in DSG Crosslinking Solution, 4mL per 10M cells. Rock
gently at room temperature for 45 minutes. NOTE: Upon addition of
crosslinkers to the pellet it is critical to pipette up and down repeatedly to
break up any clumps of cells to avoid any cell-cell crosslinking.

5. Pellet cells for 4 minutes at 1000 X G at room temperature. Discard
supernatant.

12
 SPRITE Protocol

6. Wash cells with 4mL 1x PBS per 10M cells. Pellet as before, discarding
supernatant.

7. Resuspend cell pellet in 3% formaldehyde in PBS. Rock gently at room
temperature for 10 minutes. NOTE: Upon addition of crosslinkers to the
pellet it is critical to pipette up and down repeatedly to break up any clumps
of cells to avoid any cell-cell crosslinking.

8. Add 200uL of 2.5M glycine stop solution per 1mL of cell suspension. Rock
gently at room temperature for 5 minutes.

9. Pellet cells at 4C for 4 minutes at 1000 X G at room temperature. Discard
formaldehyde supernatant in an appropriate waste container. From here,
keep cells at 4C.

10. Resuspend cell pellet in cold Scraping Buffer and gently rock for 1-2 minutes.

11. Pellet cells at 4C for 4minutes at 1000 X G. Discard supernatant in
formaldehyde waste container.

12. Resuspend cell pellet in cold Scraping Buffer again and gently rock for 1-2
minutes. Pellet as before and discard supernatant.

13. Resuspend pellet in 1mL of Scraping Buffer per 10M cells.

14. Aliquot 10M cells each into Microcentrifuge tubes and pellet at 4C for 5
minutes at 2000 X G. Remove supernatant.

15. Flash freeze in liquid nitrogen and store pellet at -80C.

(1) Engreitz, Jesse “RNA Antisense Purification (RAP): Experimental Protocols”


4.2 Cell Lysis

1. Chill Lysis Buffers A, B, and C on ice.

2. If using an electronic chiller for the sonication chamber, pre-chill to 4C.

3. Thaw 10M cell pellets on ice.

4. Add 1.4mL of Lysis Buffer A supplemented with 1x Proteinase Cocktail
Inhibitor (PIC) to each 10M cell pellet and resuspend.

5. Incubate mixtures on ice for 10 minutes.

13
 SPRITE Protocol

6. Pellet cells at 4C for 9 minutes at 850 X G.

7. Discard the supernatant, taking care not to disturb the pellet.

8. Add 1.4mL of Lysis Buffer B supplemented with 1X PIC to each 10M cell
pellet and resuspend.

9. Incubate mixtures on ice for 10 minutes.

10. Pellet cells at 4C for 9 minutes at 850 X G.

11. Discard the supernatant, taking care not to disturb the pellet.

12. Add 550uL of Lysis Buffer C supplemented with 1X PIC to each 10M nuclei
pellet and resuspend.

13. Incubate mixture on ice for 8 minutes.

14. Sonicate each sample at 4-5 watts for 1 minute: 1 pulse for 0.7 seconds ON,
3.3 seconds OFF. During and after sonication, keep lysate at 4C. A Branson
needle-tip sonicator kept at 4C was used for this protocol.

15. Pool all lysates together and split again into 10M aliquots. This ensures that
all samples in each tube are equally processed and DNAsed in the subsequent
steps.

16. Flash freeze lysate and store at -80C.


4.3 DNA Fragmentation

1. Thaw one tube of lysate on ice.

2. To determine the optimal amount of DNase to use for DNA fragmentation,
test varying DNase concentrations on 10uL aliquots of lysate.

Stock Solution Volume
10X DNase Buffer 2uL
Lysate 10uL
Turbo DNase from ThermoFisher 2 / 3 / 4 / 5 / 6 uL
H20 6 / 5 / 4 / 3 / 2 uL
Total 20uL


3. Incubate at 37C for 20 minutes.

14
 SPRITE Protocol

4. Add 1uL of 25X DNase Stop Solution to each sample to terminate the
reaction.



5. Reverse the crosslinks in each sample.

Stock Solution Volume
Lysate 21uL
MyRNK Buffer 71uL
Proteinase K 8uL
Total 100uL

6. Incubate for at 65C for three hours at the minimum, optimally overnight.

7. Follow the protocol provided in the DNA Clean and Concentrator-5 Kit,
binding in 6 volumes of DNA Binding Buffer. Elute in 10uL of H20.

8. Run each DNase sample on a gel with a 100bp DNA ladder. An ideal
fragmentation sample will have most DNA around 200bp. Size should not
greatly exceed 1kb, or be overly fragmented where DNA is <100bp. Run
traces that have no long tail of DNA fragments on a DNA HS Bioanalyzer of
D1000 Tapestation. Example trace of DNA sizes obtained for SPRITE.

9. If none of these concentrations of TURBO DNase led to ideal fragmentation,
adjust concentrations and repeat the DNasing until optimal conditions are
found.

10. DNase the batch of crosslinked lysate at the identified optimal DNAase
concentration

Stock Solution Volume
10X DNase Buffer 110uL
Lysate 550uL

15
 SPRITE Protocol

Turbo DNase from ThermoFisher X uL

H20 X uL to reach final volume
Total 1100uL


11. Incubate at 37C for 20 minutes.

12. Add 44uL of 25X DNase Stop Solution to each sample to terminate the
reaction.

13. Flash freeze DNase lysate and store at -80C.

16
 SPRITE Protocol

5 SPRITE Pre Split-and-Pool and QC Pt. 1

Goal: Lysate is coupled to Pierce NHS-Activated Magnetic Beads to allow for easy
DNA library preparation. DNA overhangs caused by fragmentation are repaired and
blunted by a combination of T4 Polynucleoide Kinase, which adds phosphate onto 5’
ends, and T4 DNA Polymerase, which has 5’ to 3’ polymerase activity as well as 3’ to
5’ exonuclease activity. Klenow fragment (-exo) is used to add an adenine dNTP to 3’
ends of each DNA molecule. This aids in ligation of the DPM adaptor, which has a 3’
thymine overhang, without creating spurious ligation products.

It is crucial to have an optimal bead to molecule ratio for the library preparation and
SPRITE processes. We aim to bind at a 1:4 to 3:4 ratio of DNA molecules to beads;
generally we bind around 50 billion molecules to 75 billion beads. Assuming that we
have 50% binding efficiency, coupling ratio is then 1:8 to 1:2.7 molecules per bead.
We assume that there are far more molecules than crosslinked complexes making
the complex to bead ratio even lower than stated above, but use molecules as an
over-estimate to reduce noise per bead. To determine the microliter amount of
lysate (DNA and/or RNA) to couple we calculate the lysate molarity by running a 5%
aliquot on the Qubit Fluorometer to determine concentration of molecules per ul
lysate and the Agilent Bioanalyzer to determine average size.

5.1 NHS Coupling

Note 1: All wash steps at 4C are performed in a cold room. All wash steps above
room temperature are performed on an Eppendorf Thermomixer. If a temperature
is not specified, it is at room temperature. To wash beads, place the tube containing
the beads on a magnetic rack to capture the beads. Wait until the solution is clear
and all beads are captured before removing the liquid. Add the wash solution to the
beads and remove the tube from the magnet. Gently pipette with a low-bind tip to
mix thoroughly until all beads are in suspension. If using an Eppendorf
Thermomixer, set the thermomixer to shake at 1200RPM. Then place the tube back
on the magnet to capture the beads again. Wait until the solution is clear and all
beads are captured before removing the wash liquid. NOTE: These steps are critical
to avoid loss of beads throughout protocol

Note 2: The protocol can be stopped at any point of the process. To ensure the
integrity of the DNA, resuspend the beads in 1mL RLT++ and store at 4C until you
wish to resume. Wash three times with M2 Buffer to remove all RLT before
proceeding with the protocol to prevent denaturating enzymes in subsequent steps
of the protocol.

Note 3: All steps involving bead pipetting should use low-bind pipette tips.

1. Gently invert the bottle containing the Pierce NHS-activated beads in N,N-
dimethylacetamide (DMAC) until there is a uniform suspension. Being careful

17
 SPRITE Protocol

not to introduce water into the bottle, transfer 2mL of NHS beads into a clean
1.7mL tube. Place the tube on a magnetic rack to capture the beads.

2. Remove the DMAC and wash beads with 1mL ice-cold 1mM HCl.

3. Wash beads with 1mL ice-cold 1X PBS.

4. Add 1mL Coupling Buffer to the beads. Before mixing, add the appropriate
amount of lysate to the coupling buffer.

5. Incubate the lysate and beads overnight at 4C on a mixer.

6. Place beads on a magnet and remove a 500uL flowthrough aliquot to another
tube. This aliquot can be analyzed to determine how much lysate was
coupled.

7. Add 500uL 1M Tris pH 7.5 (3M ethanolamine pH 9.0 can also be used), to the
beads and incubate on a mixer at 4C for at least 45 minutes. This ensures that
all NHS beads will be quenched with protein from bound lysate or Tris, and
will not bind enzymes in the following steps.

8. Wash beads four times in cold RLT++ Buffer at 4C for 3-5 minutes each time.

9. Wash beads twice in PBLSD+ Wash Buffer at 50C for 4-5 minutes each time.

10. Wash beads once at room temperature in PBLSD+ buffer.

11. Wash beads three times with M2 Buffer.

12. Spin the beads down quickly in a microcentrifuge and place back on the
magnet to remove any remaining liquid.


5.2 Phosphorylation and End Repair

1. Phosphorylate the 5’ ends of the DNA molecules to allow subsequent barcode
ligation by adding the following mixture to the beads:

Stock Solution Volume
H20 167.5uL
T4 Polynucleotide Kinase Reaction 20uL
Buffer (10X)
T4 Polynucleotide Kinase 10uL
100mM ATP 2.5uL
Total 200uL

18
 SPRITE Protocol

2. Incubate on a thermomixer for 60 minutes at 37C, 1200RPM.

3. Wash beads three times with M2 Buffer.

4. Spin the beads down quickly in a microcentrifuge and place back on the
magnet to remove any remaining liquid.

5. Blunt the 5’ and 3’ ends of the DNA molecules to prevent unwanted ligation
by adding the following mixture to the beads:

Stock Solution Volume
H20 212.5uL
End Repair Reaction Buffer (10X) 25uL
End Repair Enzyme Mix 12.5uL
Total 250uL

6. Incubate on a thermomixer for 60 minutes at 24C, 1200RPM.

7. Wash once with RLT++ Buffer.

8. Wash three times with M2 Buffer.

9. Spin the beads down quickly in a microcentrifuge and place back on the
magnet to remove any remaining liquid.

10. Add dATP to the 3’ ends of each DNA molecule to allow for ligation of the
DPM adaptor by adding the following mixture to the beads:

Stock Solution Volume
H20 215uL
dA-Tailing Reaction Buffer (10X) 25uL
Klenow Fragment (exo-) 10uL
Total 250uL

11. Incubate on a thermomixer for 60 minutes at 37C, 1200RPM. If ligating the
first adaptor barcode on the same day, set up the reaction during this
incubation.

12. Wash once with RLT++ Buffer.

13. Wash three times with M2 Buffer.

14. Spin the beads down quickly in a microcentrifuge and place back on the
magnet to remove any remaining liquid.

19
 SPRITE Protocol

5.3 DPM Adaptor Ligation

Note: There are 96 adaptors that are designed to ligate onto the DNA molecules.
These DPM adaptors are kept in a 96-well stock plate at 45uM. The ligation reaction
between the adaptors and the DNA occurs in a 96-well plate. The following steps
that detail set up are designed for optimum efficiency during the process.

Note: All ligation steps include M2 buffer, which contains detergents, to prevent
beads from aggregation of multiple beads, from sticking to the plastic tips and tubes,
and for even distribution of the beads across a 96-well plate. We have verified that
these detergents do not significantly inhibit ligation efficiency.

1. Aliquot 200uL of 2x NEB Instant Sticky End Ligase Master Mix (NOTE: we
have found that other concentrations from 0.1x-1x final may be used but
might have differences in ligation efficiency) into each well of a 12-well strip
tube. Keep on ice until ready to use.

2. Centrifuge the DPM adaptor stock plate before removing the foil seal. Aliquot
2.4uL from the stock plate of DPM adaptors to a new low-bind 96-well plate.
Be careful to ensure that there is no mixing between wells at any point of the
process to avoid cross-contamination of barcodes. Use a new pipette tip for
each well. After transfer is complete, seal both plates with a new foil seal.

3. Create a diluted M2 Buffer by mixing 1100uL of M2 Buffer with 792uL of
H20.

4. Accounting for bead volume, add the M2+H20 mix to the beads to achieve a
final volume of 1700uL. Ensure that the beads are equally suspended in the
buffer.

5. Aliquot 140uL of the bead mix into each well of a 12-well strip tube.

6. Centrifuge the 96-well plate containing the aliquoted adaptors, and then
remove the foil seal.

7. Aliquot 17.6 uL of beads into each well of the 96-well plate that contains
2.4uL of the DPM adaptors. Be careful to ensure that there is no mixing
between wells at any point of the process. Use a new pipette tip for each well.
Also be careful to ensure that there are no beads remaining in the pipette tip.

8. Carefully add any remaining beads to individual wells on the plate in 1uL
aliquots.

20
 SPRITE Protocol

9. Aliquot 20uL of Instant Sticky End Ligase Master Mix into each well, mixing
by pipetting up and down 10 times. Be careful to ensure that there is no
mixing between wells at any point of the process. Use a new pipette tip for
each well.

10. The final reaction components and volumes for each well should be as
follows:

Stock Solution Volume
Beads + M2 + H20 Mix 17.6uL
DPM Adaptor (45uM) 2.4uL
2X Instant Sticky End Ligation Master Mix 20uL
Total 40uL


11. Seal the plate with a foil seal and incubate on a thermomixer for 60 minutes
at 20C, shaking for 15 seconds at 1600RPM every minute to prevent beads
from settling to the bottom of the plate. NOTE: ligation time is critical for high
efficiency of ligation each round.

12. After incubation, centrifuge the plate before removing the foil seal.

13. Pour RLT++ Buffer into a sterile plastic reservoir, and transfer 100uL of
RLT++ into each well on the 96-well plate to stop the ligation reactions. It is
not necessary to use new tips for each well.

14. Pool all 96 stopped ligation reactions into a second sterile plastic reservoir.

15. Place a 15mL conical tube on an appropriately sized magnetic rack and
transfer the pool into the conical. Capture all beads on the magnet, disposing
all RLT++ in an appropriate waste receptacle.

16. Remove the 15mL conical containing the beads from the magnet and
resuspend beads in 1mL PBLSD+ Wash Buffer. Transfer the bead solution to
a microcentrifuge tube.

17. Wash three times with PBLSD+ Wash Buffer at 50C, 1200RPM for 3 minutes
each time.

18. Wash three times with M2 Buffer.

21
 SPRITE Protocol

5.4 Quality Control (QC): Check to Determine Ligation Efficiency of the DPM
Adaptor

1. Resuspend the beads in MyRNK Buffer so that the final beads + buffer volume
is 1mL. Remove a 5% aliquot (50uL) into a separate microcentrifuge tube.

2. Place the remaining 95% of beads back on the magnetic rack, remove the
MyRNK Buffer, and store beads in 1mL of RLT++ Buffer. Keep beads at 4C
overnight.

3. Remove the DNA+DPM adaptor molecules from the beads and reverse the
crosslinks in the lysate. Proteinase K will degrade the protein covalently
linking the the proteins in the crosslinked lysate to the NHS group on the
beads, releasing the lysate from the beads. Proteinase K and heat will further
degrade any protein in the lysate and remove crosslinks, leaving nucleic acid
in the sample.

Stock Solution Volume
Sample on beads in MyRNK Buffer 50uL
MyRNK Buffer 42uL
Proteinase K 8uL
Total 100uL

4. Incubate at 65C overnight. (Time can be reduced to 1-4hrs if needed)

5. Place the microcentrifuge tube on a magnet and capture the beads. Remove
the flowthrough that contains the DNA ligated with DPM adaptor and place in
a clean microcentrifuge tube.

6. Pipette 25uL of H20 into the tube containing the beads. Vortex, and re-
capture the beads. Remove the 25uL of H20 that now contains any residual
nucleic acid and add to the new sample tube. Discard the beads.

7. Follow the protocol provided in the DNA Clean and Concentrator-5 Kit,
binding in 6 volumes of DNA Binding Buffer. Elute in 40uL of H20.

8. Amplify the DNA molecules that are ligated to the adaptors. The forward
primer should prime off the 5’ end of the DPM adaptor and the reverse
primer should prime off the 3’ end of the DPM adaptor. Before placing the
reaction in the thermocycler, split the sample into two tubes with 50uL in
each tube.

Stock Solution Volume
Sample (cleaned) 10uL
DPMQCForward Primer (100uM) 2uL

22
 SPRITE Protocol

DPMQCReverse Primer (100uM) 2uL

H20 36uL
Q5 Hot Start Master Mix 50uL
Total 100uL

PCR Program:
1. Initial denaturation: 98C- 180 seconds
2. 12-16 cycles:
a. 98C-10 seconds
b. 68C - 30 seconds
c. 72C- 60 seconds
3. Final extension: 72C- 180 seconds
4. Hold 4C

9. Clean the PCR reaction and size select for your target DNA molecules. Our
DPM adaptors are 30 base pairs each and our target DNA molecules no less
than 100 base pairs. Agencourt AMPure XP beads size select while cleaning
the PCR reaction of unwanted products.

a. Combine the two 50uL PCR reactions back into one tube.
b. Add 1.0X AMPure XP beads to the sample for a total volume of 200uL
and mix thoroughly.
c. Incubate for 10 minutes at room temperature, mixing again at 5
minutes.
d. Place the beads on an appropriately sized magnet to capture the
beads and the bound DNA. Wait a few minutes until all the beads are
captured.
e. Remove the supernatant and discard.
f. Wash beads twice with 70% ethanol by pipetting ethanol into the tube
while beads are captured, moving the tube to the opposite side of the
magnet so that beads pass through the ethanol, and then removing the
ethanol solution.
g. Quickly spin down the beads in a microcentrifuge, re-capture on
magnet, and remove any remaining ethanol.
h. Air-dry beads while the tube is on the magnet.
i. Elute the amplified DNA from the beads by resuspending the beads in
12uL of H20. Place the solution back on the magnet to capture the
beads. Remove the eluted amplified DNA to a clean microcentrifuge
tube.

10. Run samples on a gel with a 100bp DNA ladder to confirm that DNA
molecules have ligated to the DPM adaptors. If you have enough samples, it
may be worth the cost to run samples on the Agilent BioAnalyzer to better
ascertain the library distribution and determine average peak size to
calculate molarity and molecule number. Before samples are run on the

23
 SPRITE Protocol

BioAnalyzer, their concentration is determined from the Qubit Fluorometer.

No more than 5ng of DNA is loaded into each well on the BioAnalyzer chip.
Follow instructions provided with the High Sensitivity dsDNA kit.

11. The DPM amplified libraries should reflect a similar size distribution (~200-
400bp average size after DPM adaptor ligation) that was observed during the
DNAse digestion step. If a clear library is not visible on DNA HS Bioanalyzer
or D1000 HS after 14-16 cycles, it is not advised to proceed, as the SPRITE
split-and-pool barcoding will be performed on too little material.

Example DPM library after 12 cycles on 5% of the ligated material on a DNA
HS Bioanalyzer.

Example DPM library after 16 cycles on 5% of the ligated material on 2%
agarose gel.

24
 SPRITE Protocol

6 SPRITE and Library Preparation Pt. 2

Goal: The SPRITE method provides each DNA-DNA complex in the sample lysate
with a unique nucleic acid barcode. When these complexes are decrosslinked, the
individual DNA molecules that made up a single complex retain identical barcodes.
These DNA libraries are sequenced on an Illumina Next-Generation sequencing
platform and analyzed. Any DNA molecules found to have the same barcode interact
in-vivo.

The SPRITE method works by splitting into a 96-well plate a pooled sample of
crosslinked lysate where DNA molecules are ligated to the DPM adaptor. Each well
of the 96-well plate contains a unique tag (Odd) to which the DNA molecules are
ligated. The ligation reactions are stopped, pooled, and split again into a new 96-
well plate containing different, unique tags than the first (Even). If n rounds of tag
ligation are performed, 96n unique barcodes are generated. We typically ligate 5
tags, creating over 8 billion unique barcodes. After all barcodes are ligated, the
sample is split again into small m aliquots (100 wells of 1% aliquots up to 10 wells
of 10% aliquots are typically used depending on the total material coupled) for PCR
amplification. This final splitting of samples effectually sorts the DNA complexes
once more, so that the chance that two different non-crosslinked complexes with the
same barcode are amplified together is negligible. This last dilution into m wells
effectively raises the number of unique tags to each molecule to m*96n. For example,
if the sample is aliquoted into 1% aliquots, then over 815 billion unique barcodes
are generated.

The first round of SPRITE was already completed with the ligation of 96 unique
DPM adaptors that allow for the subsequent ligation of new barcodes. As detailed in
the Adaptor and Barcode Creation section, subsequent tag ligations are performed
in the following order:

1. ODD Tag Ligation
2. EVEN Tag Ligation
3. ODD Tag Ligation
4. Terminal Tag Ligation

The four barcode ligations listed above are performed in the exact same manner
with the only difference being the tag sequence. Thus, the following section will only
detail one round of SPRITE.


6.1 SPRITE

1. Aliquot 200uL of Instant Sticky End Ligase Master Mix into each well of a 12-
well strip tube. Keep on ice until ready to use.

25
 SPRITE Protocol

2. Centrifuge the tag stock plate before removing the foil seal. Aliquot 2.4uL
from the stock plate of barcodes to a new low-bind 96-well plate. Be careful
to ensure that there is no mixing between wells at any point of the process.
Use a new pipette tip for each well. After transfer is complete, seal both
plates with a new foil seal.

3. Create a diluted M2 Buffer by mixing 1100uL of M2 Buffer with 792uL of
H20.

4. Accounting for bead volume, add the M2+H20 mix to the beads to achieve a
final volume of 1700uL. Ensure that the beads are equally suspended in the
buffer.

5. Aliquot 140uL of the bead mix into each well of a 12-well strip tube.

6. Centrifuge the 96-well plate containing the aliquoted barcodes, and then
remove the foil seal.

7. Aliquot 17.6 uL of beads into each well of the 96-well plate that contains
2.4uL of the tags. Be careful to ensure that there is no mixing between wells
at any point of the process. Use a new pipette tip for each well. Also be
careful to ensure that there are no beads remaining in the pipette tip.

8. Carefully add any remaining beads to individual wells on the plate in 1uL
aliquots.

9. Aliquot 20uL of Instant Sticky End Ligase Master Mix into each well, mixing
by pipetting up and down 10 times. Be careful to ensure that there is no
mixing between wells at any point of the process. Use a new pipette tip for
each well.

10. The final reaction components and volumes for each well should be as
follows:

Stock Solution Volume
Beads + M2 + H20 Mix 17.6uL
Tag (45uM) 2.4uL
2X Instant Sticky End Ligation Master 20uL
Mix
Total 40uL


11. Seal the plate with a foil seal and incubate on a thermomixer for 60 minutes
at 20C, shaking for 15 seconds at 1600RPM every 5 minutes.

26
 SPRITE Protocol

12. After incubation, centrifuge the plate before removing the foil seal.

13. Pour RLT++ Buffer into a sterile plastic reservoir, and transfer 100uL of
RLT++ into each well on the 96-well plate to stop the ligation reactions. It is
not necessary to use new tips for each well.

14. Pool all 96 stopped ligation reactions into a second sterile plastic reservoir.

15. Place a 15mL conical tube on an appropriately sized magnetic rack and
transfer the pool into the conical. Capture all beads on the magnet, disposing
all RLT++ in an appropriate waste receptacle.

16. Remove the 15mL conical containing the beads from the magnet and
resuspend beads in 1mL PBLSD+ Wash Buffer. Transfer the bead solution to
a microcentrifuge tube.

17. Wash three times with PBLSD+ Wash Buffer at 50C, 1200RPM for 3 minutes
each time.

18. Wash three times with M2 Buffer.

19. Repeat the process starting at Step 1 for the remaining three or more SPRITE
rounds.


6.2 Library Preparation Pt.2

1. Resuspend the beads in MyRNK Buffer so that the final beads + buffer volume
is 1mL.

2. Remove five aliquots into clean microcentrifuge tubes: 0.5%, 1%, 2.5%, 5%,
and 7.5% (5uL, 10uL, 25uL, 50uL, and 75uL) and elute the barcoded DNA
from the beads.

Stock Solution Volume
Sample on beads in MyRNK Buffer 5 / 10 / 25 / 50 / 75uL
MyRNK Buffer 87 / 82 / 67 / 42 / 17uL
Proteinase K 8uL
Total 100uL


3. Incubate at 65C overnight.

27
 SPRITE Protocol

4. Place the microcentrifuge tubes on a magnet and capture the beads. Remove
the flowthrough that contains the barcoded DNA and place in a clean
microcentrifuge tube.

5. Pipette 25uL of H20 into the tube containing the beads. Vortex, and re-
capture the beads. Remove the 25uL of H20 that now contains any residual
nucleic acid and add to the new sample tube. Discard the beads.

6. Follow the protocol provided in the DNA Clean and Concentrator-5 Kit,
binding in 6 volumes of DNA Binding Buffer. Elute in 40uL of H20.

7. Amplify the final barcoded DNA through PCR. Refer to section 3.4 for dtails
about the final library amplification step. Before placing the reaction in the
thermocycler, split the sample in in to two tubes with 50uL in each tube.

Stock Solution Volume
Sample (cleaned) 40uL
First Primer (100uM) 2uL
Second Primer (100uM) 2uL
H20 6uL
Q5 Hot Start Master Mix 50uL
Total 100uL

PCR Program:
1. Initial denaturation: 98C- 180 seconds
2. 4 cycles:
a. 98C-10 seconds
b. First Annealing Temperature- 30 seconds
c. 72C- 90 seconds
3. 5 cycles:
a. 98C-10 seconds
b. Second Annealing Temperature- 30 seconds
c. 72C- 90 seconds
4. Final extension: 72C- 180 seconds
5. Hold 4C

8. Clean the PCR reaction and size select for your target libraries. The total
length of our barcode on one amplified product is around 160 base pairs and
each target DNA molecules no less than 100 base pairs. Agencourt AMPure
XP beads are able to size select while cleaning the PCR reaction of unwanted
products.

a. Combine the two 50uL PCR reactions back into one tube.
b. Add 0.7X AMPure XP beads to the sample for a total volume of 170uL
and mix thoroughly.

28
 SPRITE Protocol

c. Incubate for 10 minutes at room temperature, mixing again at 5

minutes.
d. Place the beads on an appropriately sized magnet to capture the
beads and the bound DNA. Wait a few minutes until all the beads are
captured.
e. Remove the supernatant and discard.
f. Wash beads twice with 70% ethanol by pipetting ethanol into the tube
while beads are captured, moving the tube to the opposite side of the
magnet so that beads pass through the ethanol, and then removing the
ethanol solution.
g. Quickly spin down the beads in a microcentrifuge, re-capture on
magnet, and remove any remaining ethanol.
h. Air-dry beads while the tube is on the magnet.
i. Elute the amplified DNA from the beads by resuspending the beads in
100uL of H20. Place the solution back on the magnet to capture the
beads. Remove the eluted amplified DNA to a clean microcentrifuge
tube.
j. Repeat the clean up with 0.7X AMPure XP beads, eluting finally in
12uL.

9. Determine the concentration of each library with the Qubit Fluorometer. Our
final libraries are generally between 0.5ng/uL and 1.5ng/uL.

10. Load all samples on the Agilent BioAnalyzer, following the protocol provided
with Agilent’s High Sensitivity dsDNA Kit. Final library sizes range from
around 260 base pairs to 1000 base pairs with peaks around 400 base pairs.

11. Using the concentrations gathered from Qubit and the average library size
gathered from the BioAnalyzer, estimate the number of DNA molecules in
each library. These numbers are used to determine the depth to sequence
each library (see 6.3).

a. Formula to calculate Library Concentration nM post-pcr:
i. (Concentration ng/ul)/((649*103*Average Size bp))*109
b. Formula to calculate molecules post-PCR:
i. (10-15 *Lib concentration nM)*(Vol. eluted uL)*(6.022*1023)
c. Formula to calculate molecules pre-pcr:
i. NOTE: This has been adjusted to account for loss during 2
rounds of 0.7x SPRI and depends on the number of cycles PCR
amplified.
ii. (Molecules post-PCR)/(2(cycles+2))

6.3 Estimating sequencing depth

SPRITE interactions are defined based on the sequences that share the same tags.
Accordingly, it is essential to sequence all of the barcoded molecules in a complex in

29
 SPRITE Protocol

order to identify interactions in a sample. Therefore, the number of unique

molecules that are sequenced dramatically affects the likelihood of identifying
interacting molecules. To address this, we optimized the loading density of our
sequencing sample based on the number of unique molecules contained in the
sample. Our goal is to load approximately equimolar unique molecules as the
number of sequencing reads generated. Specifically, based on our simulations of
Poisson sampling, we have found that sequencing with ~1-3x coverage of reads per
the number of unique molecules will ensure that most molecules are sampled. This
follows Poisson sampling where 1-1/ec of molecules are sampled at a given c
coverage. For example, 3x, 2x, and 1x coverage samples approximately 95%, 86%,
and 63% of interactions, respectively. In this study, most libraries were sampled
with approximately 1.5-2x coverage.

To determine the number of unique molecules in our sample, we measure the
amount of material present on beads prior to reverse crosslinking all interactions.
To do this, we take an aliquot of the sample and reverse crosslink to elute (as
above), cleanup DNA, and PCR amplify for 9-12 cycles. We then measure the
molarity using the Qubit and Bioanalyzer (as above). The number of unique
molecules prior to PCR is back calculated from a standard curve and adjusted to
account for loss during the cleanup. Specifically, we calculate the number of
molecules pre-pcr (see above). This is used to estimate the number of unique
molecules in the sample prior to PCR. For example, if we estimate there are 70M
unique molecules pre-pcr in a library, we will sample that library with 105-140M
reads. If the number of molecules pre-pcr exceeds the number of molecules that can
be sampled on a sequencer (~300-400M) you must take a smaller aliquot and ProK
an aliquot with a smaller amount of material or you will not detect all interactions in
a sample and many molecules will be non-interacting.

In addition to optimizing molarity, because this dilution results in approximately
1% aliquots of the total sample being separately eluted and amplified, this
effectively serves as another round of split-pool barcoding as each library is tagged
with a unique barcoded Illumina primer. This further reduces the probability that
molecules in different clusters obtain the same barcodes.

30
 SPRITE Protocol

7 Sequencing and Data Analysis

The Illumina, Inc. HiSeq v2500 platform was employed for next generation
sequencing of the generated libraries using a TruSeq Rapid SBS v1 Kit – HS (200
cycle) and TruSeq Rapid Paired End Cluster Kit – HS. All SPRITE data in this paper
was generated using Illumina paired-end sequencing. Reads must be long enough to
incorporate all tag information. Most read-pairs in this report were (115 bp, 100
bp).

All code for the SPRITE computational pipeline is found here:
https://github.com/GuttmanLab/sprite-pipeline/wiki

7.1 Tag identification

This step is performed using custom in-house software. The program takes as input
both FASTQ files, sorted by name so that the record with a particular line number in
the read 1 file corresponds with the record with the same line number in the read 2
file. The program also requires a text file containing the tag sequences with unique
identifiers and an identification tolerance -- the number of mismatches tolerated
between the tag and the read when search for the tag.

The program first loads the tags from the tag file and stores them in a hashtable
keyed by sequence. Storing these sequences in a hashtable allows rapid (O(1))
string matching. Additional tags are generated according to the given identification
tolerances, and these are also stored. For example, if the tag TTTT has an
identification tolerance of 1, the tag will be inserted into the table, keyed by all
sequences at most one Hamming distance away:

TTTT
ATTT
TATT
TTAT
TTTA
CTTT
TCTT
TTCT
TTTC
GTTT
TGTT
TTGT
TTTG
NTTT
TNTT
TTNT
TTTN

After storing the tags, the program iterates through the read-pairs by advancing
line-by-line through both FASTQ files simultaneously. For a given sequence, the

31
 SPRITE Protocol

program queries the hash table for substrings that correspond to known tag
positions. (The exact details of this process depend on the barcoding scheme.) After
the identification process for a record is complete, the tags are appended to the
name of the record, and this modified record is output into new read 1 and read 2
FASTQ files.

7.2 Alignment

In our barcoding schemes, only one of the reads in a read pair contains an
appreciable amount of genomic sequence. These genomic-reads are aligned to the
appropriate reference with Bowtie2 under the default parameters — except for the
following.

Only one of the two FASTQ files is aligned. We do not run a paired-end alignment
despite having paired-end reads.

Before the genomic sequence on the read is an 11-mer DPM tag sequence. To
account for this, we run Bowtie2 with `--trim5 11`.

After the sequence, there are two possibilities. The read may extend into the tag
sequences on the other end of the fragment if the fragment is too short, or the read
may terminate before the tags if the fragment is long enough. To account for the
inclusion of tag sequences, we run Bowtie2 with `--local`. (This would also deal with
the DPM tag at the start of the sequence.

We align to both the reference chromosomes and unplaced scaffolds (typically end
in “random”).

We sort the resulting SAM file and convert it to a BAM file. The names of each SAM
record contain the identified tags, as these were present in the input FASTQ files.

7.3 Filtration

The BAM file is then passed through successive filtration steps:

i. Remove all alignments with a MAPQ score less than 30. This removes all
unmapped reads. Note that the MAPQ score depends on the aligner used; it is
not standardized. If a different aligner is used, this step will need to be
replaced with a different quality-filtration step.
ii. Remove all alignments that align to the reference with a Hamming score > 2.
We only tolerate two mismatches at most between the read and the
reference.
iii. Remove all alignments that overlap (in any amount) any region in the repeat-
mask BED file provided to us by B. Tabak. We used bedtools intersect with
the `-v` flag set.

32
 SPRITE Protocol

iv. Remove all alignments that overlap (in any amount) any region in the mask
BED file generated by ComputeGenomeMask in the GATK package from the
Broad. This mask file was generated by shredding the reference into 35-mers
and BLATting them against the reference. Any non-unique location that a 35-
mer maps to is masked. The output of ComputeGenomeMask is not a BED file,
but a FASTA file where all masked bases are represented with 0s, and all
unmasked bases are represented with 1s. This mask file is converted to a
BED file with a custom Python script.

7.4 Subsequence post-processing

See the Guttman Lab Github page for the post-processing scripts to identify
interactions and QC the library is found here. https://github.com/GuttmanLab/barcoding-
post/wiki

7.5 Quality Controls of Successful SPRITE Libraries

We use the following metrics to evaluate whether SPRITE tagging was successful on
the post-sequencing library:

Calculate percentage of reads with all tags ligated (get_ligation_efficiency.py)
1. The percentage of reads with all tags ligated should be >90% each round. For
example, if r=5 rounds of SPRITE were performed, we can determine this by
calculating the fraction of reads f with all 5 barcodes and calculating the
ligation efficiency each round. Ligation efficiency per round = f 1/r. This is an
example of the distribution of reads with the number of barcodes identified.

2. Percentage of chromatin that is interacting with other chromatin: We have
found that over-fragmentation of the lysate via sonication for 15-20 minute
results in the majority of SPRITE molecules that are non-interacting.
Specifically, the majority of reads after heavy sonication do not share
barcodes with other molecules. Spinning the chromatin after 1-2 minute at a
high speed after sonication and only performing SPRITE on the supernatant
also results in a similar problem.

33
 SPRITE Protocol

3. FastQC to QC quality of reads on sequencer. Due to a monotemplate that can
occur because all reads share a sticky end, we QC whether any monotemplate
issues happen to impact the quality of the reads when a common sequence is
reached on the machine. We have obtained high quality libraries on HiSeq
2500 and HiSeq 4000 and NextSeq loading at 8pM.
4. Calculate sufficient sampling of library to ensure all tagged molecules are
sequenced, which ensures that the majority of molecules in a complex are
identified on the sequencer. We use the program PreSeq from the Smith Lab
to ensure that most of the unique molecules (>70%) estimated by the
program in a library are sampled. We typically sample at a depth of 1.5-2x
reads per the total number of unique molecules estimated in a library. For
example, if we estimate there are 70M unique molecules in a library, we will
sample that library with 105-140M reads. Sequencing at higher depth results
in many duplicates without gaining many more molecules in each complex.
Highly under-sampling (<40%) however results in a lot of molecules that are
not interacting with any other molecule and need to be sequenced further.

34