Supplemental Figure 1. Overview of Main Chip-Seq Procedures Pioneered for Low-Input And/Or Automated Chip-Seq

Supplemental "gures “A plug and play micro!uidic platform for standardized sensitive low-input Chromatin Immunoprecipitation”, Dirks et al., Genome Research 2021 Supplemental Figure 1

Supplemental Figure 1. Overview of main ChIP-seq procedures pioneered for low-input and/or automated ChIP-seq. In green the features that are advantageous, in red the disadvantageous features, in orange strategies that are compatible with single cell readout. Current automated low-input ChIP-seq work!ows require custom-built platforms and can handle only a low number of parallel samples, as also indicated in the red boxes. Supplemental Figure 2

Chromatin Immunoprecipitation-sequencing work!ow

Collection of biomaterial

Chromatin isolation and DNA cleavage (digestion or sonication)

Automation and parallelization on micro!uidics platform 1) Load reagents Antibody coupling

Antibody-complex binding to sca#old

Multi-step washing 2) Run protocol

DNA release

DNA puri"cation

Library construction A C T G . . . and sequencing A A G C . . . C G T A . . .

Supplemental Figure 2. Overview of the conventional ChIP-seq work!ow and the part of the ChIP-seq work- !ow that we automated on the micro!uidic platform. Supplemental Figure 3 Micro!uidic ChIP pipette map A

C1 Accumulator C2

X W W X W X V U U V U V T S S T S T R Q Q R Q R P O O P O P N M M N M N K L L K L K PDMS Beads inputs Beads Harvest outlets Harvest I J circuitry Inputs Sample J I J I G H H G H G E F F E F E C D D C D C A B B A B A 5 3 2 1 C3 C4 6 4 Accumulator

Carrier inlet Buer C1, C2, C3, C4 25µL 0.05% Tween-20 Accumulators 200µL 0.05% Tween-20 Beads inlet 15µL ProA/G beads 1 20µL Frit beads 2 25µL Equilibra buer 3 25µL Equilibra buer 4 200µL High Salt wash buer 5 5µL DNA Extrac buer 6 200 µL DNA Elu buer Sample inlet 1-8µL chroma + anbody Harvest outlet ~ 3uL output B Bead path Ab+chrom. path Washing path

23 nL volume bead reactor 15 nL downstream Supplemental Figure 3. Details of the micro!uidic IFC. ( A) Pipette map on the newly developed micro!uidic plates for ChIP. All control valves as present in the Integrated Fluidic Circuitries (Fig. 1C) can be individually pressurized by the use of 0.05% Tween 20 solution that is loaded within the wells C1-C4 and within the accumulators. The PDMS circuitry chip is present in the center. (B) Volumes of the main parts of a single reactor unit within the micro!uidic IFC. Supplemental Figure 4 A

Column size: Small Medium Large Approximate volume: 1 nL 3 nL 6-9 nL

Columns A-F

Columns G-L

Columns M-R

Columns S-X

50 45 C 40 35 30 25 20 % recovery 15 Actb 10 Neg 5 0 ABCDEFGHIJKLMNOPABCDEFGHIJKLMNOP 3000 cell microuidic ChIP - run 1 3000 cell microuidic ChIP - run 2 Supplemental Figure 4. Overview and reproducibility of the antibody binding columns as generated within the micro!uidic chip for ChIP. (A) Phase-contrast image of micro!uidic bead columns that are packed to various sizes. (B) Reproducibility of bead packing column across the 24 parallel reactors of the Integrated Fluidic Circuit. These are not the exact but representative columns for the experiments as shown in Fig. 2F and Supplemental Fig. S4C. (C) Reproducibility of ChIP-qPCR across the 24 parallel reactors of a single micro!uidic plate (A-P), and between 2 microfluidic plates run on different days (run 1 and 2), as shown for H3K4me3 per individual ChIP. Supplemental Figure 5

A Carrier inlet Hands-on me Acquire chroman, fresh or frozen, crosslinked sonicated or nave digested. -- Combine chroman with ChIP bur with anbody of choice 10 minutes Pre- -- Pipee beads, reagents and chroman into carrier 10 minutes -- Harvesng 10 minutes Total 30 minutes

Step Details Time B Load microuidic chip into the controller, pressurize the valves and inlets, priming of 15 min. valves and reagents to avoid air bubbles in the system. Primes branch structure and reactors, primes column bypass, packs 5 cycles of ~200 pL frit 30 min. beads into the reactor, packs ProAG beads into reactor for 10 minutes 5 nL . Removes bead buer from ChIP column. 10 min. Loads chroma and anbody across the bead column at the specied pressure for the 35 min. specied amount of me. Equilibrates the ChIP to physiological salt condion, then performs a high salt wash to 20 min. remove non-speci binding proteins and DNA segments. Removes the high salt wash buer and any remaining non-speci DNA carrying proteins. 10 min. Loads a Proteinase K buer on to the column and increases the temperature to elute min. speci DNA from the column. Pushes the speci DNA out of the PDMS circuitry into the carrier for pipeng o-chip. 60 min.

Script Sengs, oponal tweaks C def script.requires LIBRARY Load library containing valve operaon schemacs Miniaturized ChIP- script.version

def main ChIP_Names NGS_Names AP=30 Anvil pressure load_chip 1

ChIPColumn 2 5x0.05nL frit bead ProAG packing for 10 min AP=35 DilaonWash 3 Washing cycles, temperature DirectIP set_temp DilaonWash 4 Air purge, 20 cycles of 5nL, 12 psi DilaonWash 5 150mM NaCl wash, 15 cycles of 5nL, 12 psi DilaonWash 6 450mM NaCl wash, 15 cycles of 5nL, 12 psi DilaonWash 7 Air purge, 20 cycles of 5nL, 12 psi set_temp AP=40 Extracon False 8 AP=30 set_temp HarvestChIP LockChip Output chip

Supplemental Figure 5. Overview of the automated micro!uidic protocol that we developed. (A) Hands-on and machine time of the newly developed protocol. (B) Overview of the running time of the individual steps that are performed during the ChIP on the plates. (C) Overview of the adaptable script associated with running of the automated micro!uidic for ChIP. Pink numbers refer to panel (B). Supplemental Figure 6

2 Mb mm9 24,000,000 24,500,000 25,000,000 25,500,000 26,000,000 26,500,000 27,000,000 27,500,000 28,000,000 28,500,000 29,000,000 29,500,000 A Vmn2r114 Mmp25 Dcpp3 Ntn3 Dci Npw Hn1l Gnptg Sox8 Pigq Luc7l Atp6v0e Cuta Gm10505 Rps10 Anks1 Tulp1 Mapk14 Srsf3 Pi16 Vmn2r115 Thoc6 Sbp Ccnf Pgp Rps2 Telo2 Tekt4 Msln Decr2 Dusp1 Bnip1 Ip6k3 Grm4 Snrpc Zfp523 Gm749 Pxt1 Ppil1 Vmn2r116 Cldn6 Sbpl Abca3 Gfer Ift140 Sstr5 Narfl Nme4 Ergic1 Phf1 Hmga1 Taf11 Def6 Clps Kctd20 Mtch1 Vmn2r117 Srrm2 Amdhd2 E4f1 Tbl3 Ptx4 Lmf1 Solh Neurl1b Nkx2-5 Lemd2 Nudt3 Tcp11 Fkbp5 Brpf3 Rab44 Zfp213 Dcpp1 Abca17 Tsc2 Eme2 Unkl Gng13 Tmem8 A930001N09Rik Bak1 AI413582 Scube3 Lhfpl5 Stk38 Fgd2 Zfp13 Dcpp2 Rnps1 Noxo1 Tmem204 Tpsb2 Chtf18 Mrpl28 Kifc5b Pacsin1 Ppard Srpk1 Cdkn1a Tmem217 Zscan10 Prss30 Dnase1l2 Rpl3l Clcn7 Rpusd1 Axin1 Zbtb9 Spdef Ppard Slc26a8 Cpne5 Ccdc64b Prss22 Mlst8 Fahd1 Baiap3 Mslnl Pdia2 Ggnbp1 D17Wsu92e Fance Mapk13 BC004004 Hcfc1r1 Prss27 Caskin1 Hagh Ube2i Haghl Arhgdig Itpr3 Uhrf1bp1 Rpl10a Pnpla1 Tbc1d22b Tnfrsf12a Kctd5 Traf7 Igfals Prss34 Ccdc78 Rgs11 2900010M23Rik Tead3 1110021J02Rik Cldn9 Pdpk1 Rab26 Nubp2 Prss28 Fam173a Itfg3 4930511I11Rik Pkmyt1 Atp6v0c Pkd1 Spsb3 Prss29 Metrn 4930539E08Rik Paqr4 Tbc1d24 Nthl1 Nme3 Tpsab1 Fbxl16 Kremen2 9930021D14Rik Mrps34 Tpsg1 Wdr24 Flywch1 Slc9a3r2 Ccdc154 Jmjd8 Flywch2 Zfp598 BC003965 Stub1 Tceb2 Syngr3 Cacna1h Rhbdl1 Prss33 Rnf151 Rhot2 Prss41 Ndufb10 Wdr90 Prss32 Sepx1 Fam195a Prss21 Hs3st6 0610011F06Rik 4930528F23Rik Wfikkn1 Mapk8ip3 Rab40c Cramp1l Nhlrc4 Rab11fip3 96.4 -

Reference 0 Replicate 145.8 - 0 _ 3000 Replicate 250.5 - H3K4me3 0 _ Replicate 346.4 - 0 1000 49.9 - 0 _ 500 48.9 - 0 _ Replicate 111.7 - 0 Reference Replicate 217.1 - 0 _ Replicate 110.1 - 0 H3K4me1 Replicate 2 8.2 - 3000 0 _ Replicate 311.6 - 0 21.2 -

1000 0 _ Replicate 139.2 - 0 _ Reference Replicate 287.4 - 0 _ Replicate 1 27.2 - 0 _ H3K27ac Replicate 229.5 - 3000 0 _ Replicate 329.6 - 0 42.3 -

1000 0 _ Replicate 126.1 - 0 _ Reference Replicate 222.5 - 0 _ H3K36me3 3000 16.6 - 0 _ 1000 12.8 - 0 _

B HK4me3 ChIP-seq 2 million cell bench ChIP 3000 cells rep1 3000 cells rep2 3000 cells rep3 1000 cells 500 cells Read count Per Million mapped reads Read count Per 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 5'End 3'End 2000 4000

Supplemental Figure 6. PnP-ChIP-seq using small quantites of bulk-sonicated crosslinked chromatin. (A) Genome browser view of a 4Mb locus for PnP-ChIP-seq of H3K4me3, H3K4me1, H3K27ac and H3K36me3. (B) Average pro"le of H3K4me3 over all H3K4me3 peaks of pro"les generated by PnP-ChIP-seq using small quantites of bulk-sonicated crosslinked chromatin. The start and end of the peaks are indicated with 5’end and 3’end, respectively. Supplemental Figure 7 A

Scale 1 Mb mm9 Chr17: 24,500,000 25,000,000 25,500,000 Prss27 Amdhd2 Abca17 Dci Rab26 Nthl1 Rps2 Fahd1 Hn1l Telo2 Unkl Tpsab1 Kctd5 Ntn3 Abca3 Pgp Pkd1 Npw Hs3st6 Hagh Ift140 Gnptg Tpsb2 Pdpk1 Ccnf Rnps1 Traf7 Tsc2 Tbl3 Igfals Tmem204 Baiap3 Tpsg1 Mslnl Stub1 Rab11fip3 Atp6v0c Dnase1l2 Slc9a3r2 Sepx1 Nubp2 Ptx4 Ube2i Narfl Rab40c Tbc1d24 E4f1 Zfp598 Spsb3 Clcn7 Prss34 Haghl 1600002H07Rik Mlst8 Syngr3 Eme2 Ccdc154 Prss28 Ccdc78 9930021D14Rik Gfer Mrps34 BC003965 Prss29 Fam173a Caskin1 Noxo1 Nme3 0610007P22Rik Cacna1h Cells used Metrn Rnf151 Mapk8ip3 Rhbdl1 Ndufb10 Cramp1l in H3K4me3 CellsWdr90 used for Rpl3l Fam195a 4930528F23Rik 0610011F06Riksonication ChIP Wfikkn1 80 -

0 _ 2 million 5 million 80 - 0 _ 3000 Replicate 1 80 - 5 million 0 _ 3000 Replicate 2 80 - 0 _ 2400 80 - 60,000 1200 0 _ 80 -

0 _ 2400 80 - 30,000 1200 0 _ 80 -

0 _ 1200 80 - 15,000 600 0 _

B HK4me3 ChIP-seq

5 2 million cell bench ChIP Bench 3000 cells - sonication of 5M cells 3000 cells - sonication of 5M cells 2400 cells - sonication of 60k cells 1200 cells - sonication of 60k cells Micro!uidic 2400 cells - sonication of 30k cells 4 1200 cells - sonication of 30k cells 1200 cells - sonication of 15k cells

2.5 3.0 3.5 600 cells - sonication of 15k cells 3 2.0 2 Read count Per Million mapped reads Read count Per Read count Per Million mapped reads Read count Per 1 0 0.0 0.5 1.0 1.5 5'End 3'End 2000 4000

Supplemental Figure 7. Use of low-volume sonication on low numbers of mESCs for PnP-ChIP-seq. (A) Genome browser view of a gene-rich locus for PnP-ChIP-seq of H3K4me3 using a series of mESC input quantities for sonication. (B) Average pro"le of H3K4me3 over all H3K4me3 peaks of pro"les generated by PnP-ChIP-seq using a series of mESC input quantities for sonication. The start and end of the peaks are indicated with 5’end and 3’end, respectively. Supplemental Figure 8 A D B

Read count Per Million mapped reads H3K36me3 H3K27ac H3K4me1 H3K4me3 255 Read count Per Million mapped reads 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 1 2 3 4 5 6 4548 962 3000 cells H3K4me3 25312 5342 143 100 500 1000 3000 Reference 3000 Reference 3000 Reference 3000 Reference 1346 'n 'n 004000 2000 3'End 5'End Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 3 Replicate 2 Replicate 1 Replicate 3 Replicate 2 Replicate 1 179 146.3 _ 14.5 _ 14.9 _ 13.9 _ 45.0 _ 53.8 _ 13.3 _ 14.3 _ 58.6 _ 33.2 _ 45.1 _ 52.8 _ 78.9 _ 60.8 _ 63.5 _ 58.1 _ 36.1 _ Skap1 17.3 _ 59.3_ 46.2_ 5.9 _ 7.1 _ 0 5620 205 0 0 0 0 0_ 0 0 _ 0 _ 0 _ 0 0 _ 0 0 0 _ 0 0_ 0 _ 0_ 0 _ 0 0 _ 0 _ 100 cells 100 cells 500 cells 500 cells 1000 cells 1000 cells 1000 cells 3000 cells 3000 cells 3000 cells 2 millioncelllbenchChIP HK4me3 ChIP-seq 1000 cells H3K4me3 Snx11 Cbx1 Nfe2l1 20624 423 Copz2 1951 Cdk5rap3 Prr15l Pnpo Sp2 Mrpl10 Lrrc46 Scrn2 Osbpl7 Sp6 97,000,000 Tbx21 Tbkbp1 3308 Kpnb1 Npepps !uidic Micro Mrpl45 Gpr179 Bench Socs7 Arhgap23 044405 1004 E130012A19Rik Srcin1 1 Mb 97,500,000 C Mouse July2007(NCBI37/mm9)Chr11:96,448,723-99,329,260(2,880,538bp)

Read count Per Million mapped reads 1700001P01Rik Read count Per Million mapped reads Mllt6 Pip4k2b Cisd3 Pcgf2 Psmb3

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 B230217C12Rik 500 cells H3K4me3 16162 0 1 2 3 4 5 Cwc25 Rpl23 Lasp1 Fbxo47 Plxdc1 Cacnb1 Arl5c Rpl19 Stac2 98,000,000 Fbxl20 Med1 Cdk12 Neurod2 1810046J19Rik Ppp1r1b Stard3 'n 'n 004000 2000 3'End 5'End 309 Pgap3 Pnmt Tcap Erbb2 Grb7 Ikzf3 Zpbp2 98,500,000 Ormdl3 100 cells H3K4me3 13 5660 11731 Gsdma3 Gm12 Gsdma2 mm9 Gsdma Psmd3 Med24 600cells-sonic.of15k 100 cells 100 cells 500 cells 500 cells 1000 cells 1000 cells 1000 cells 3000 cells 3000 cells 3000 cells 2 millioncelllbenchChIP HK4me3 ChIP-seq Csf3 1200 cells-sonic.of15k 1200 cells-sonic.of30k 2400 cells-sonic.of30k 1200 cells-sonic.of60k 2400 cells-sonic.of60k Thra Nr1d1 Rapgefl1 Casc3 Msl1 Wipf2 Cdc6 Rara Top2a Gjd3 Igfbp4 99,000,000 Tns4 Ccr7 Smarce1 Replicate 3 Replicate 2 Replicate 1 Krt222 Krt24 Krt25 Krt26 Krt27 Krt28 Krt10 Krt12 Krt20 (sonication) ChIP !uidic Micro (MNase) ChIP !uidic Micro Bench ChIP Bench Supplemental Figure 8. PnP-ChIP-seq using small cell quantities by the use of MNase shearing on 15,000 mESCs. (A) Genome browser view of a 3Mb locus for PnP-ChIP-seq of H3K4me3, H3K4me1, H3K27ac and H3K36me3. (B) Average pro"le of H3K4me3 over all H3K4me3 peaks of pro"les generated by PnP-ChIP-seq using small cell quantities by the use of MNase shearing on 15,000 mESCs. The start and end of the peaks are indicated with 5’end and 3’end, respectively. (C) A comparison of H3K4me3 over all H3K4me3 peaks of pro"les generated by PnP-ChIP-seq using small cell quantities by the use of sonication or MNase shearing. The start and end of the peaks are indicated with 5’end and 3’end, respectively. (D) Overlap between de novo H3K4me3 peak calls of replicate PnP-ChIP-seq using small cell quantities by the use of MNase shearing on 15,000 mESCs. Supplemental Figure 9 A C H3K4me3 pro" indicated are peaks the of 3’end,5’endrespectively. end and with and start The colors). other (in mESCs on performed methods ChIP-seq input pro" H3K4me3 using or merged”) “ChIP called blue (in mESCs 15,000 on pro" of peaks H3K4me3 pro" H3K4me3 05 o a atmtd micro! automated an or 2015) al., al. et et Shen (Brind'Amourby developed by developed methods ChIP-seq input 9. Figure Supplemental 4219010 681963013 15956 4668 1506 16900 8412 Dirks 3k-Shen1k Dirks 3k-Brind’Amour5k Dirks 3k-Brind’Amour10k H3K4me3 ChIP-seq 51 09 524 10297 15015 Reference (2 million cell bulk experiment) 407 8356 16956 Dirks 3000 cells rep1 Dirks 3000 cells rep2 Dirks 3000 cells rep3 les generated using alternative low-cell input methods for ChIP-seq. les generated using alternative low-cell input ChIP-seq. input low-cell alternative using generated les Dirks 1000 cells rep1 Dirks 1000 cells rep2 Dirks 1000 cells rep3 Dirks 500 cells rep1 les generated by PnP-ChIP-seq using small cell quantities by the use of MNase shearing MNase of use the by quantities cell small using PnP-ChIP-seq by generated les

( Dirks 500 cells rep2

(A) 2015) Dirks 100 cells rep1

oprsn ewe Maebsd n-hPsq n atraie low-cell alternative and PnP-ChIP-seq MNase-based between Comparison Dirks 100 cells rep2

Brind’Amour 100.000 cells 308 10513 10111 Dirks 1k-Brind’Amour10k Dirks 1k-Brind’Amour5k Dirks 1k-Shen 20 69144 8619 12005 Brind’Amour 10.000 cells (C)

Intersections between de novo peak calls of H3K4me3 PnP-ChIP-seq and PnP-ChIP-seq H3K4me3 of calls peak novo de between Intersections Brind’Amour 5000 cells

Cross-correlation between H3K4me3 PnP-ChIP-seq (labeled “Dirks”) and (labeled “Dirks”) PnP-ChIP-seq H3K4me3 between Cross-correlation Shen 1000 cells rep1 Shen 1000 cells rep2 Shen 1000cellsrep2 Shen 1000cellsrep1 Brind’Amour 5000cells Brind’Amour 10.000cells Brind’Amour 100.000cells Dirks 100cellsrep2 Dirks 100cellsrep1 Dirks 500cellsrep2 Dirks 500cellsrep1 Dirks 1000cellsrep3 Dirks 1000cellsrep2 Dirks 1000cellsrep1 Dirks 3000cellsrep3 Dirks 3000cellsrep2 Dirks 3000cellsrep1 Reference (2 millioncellbulkexperiment) B

ReadRead count count ReadPer Per count Million Per Million Million mapped mapped mapped reads reads reads 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.00.0 0.50.5 1.0 1.5 1.0 2.0 2.51.5 3.0 3.52.0 Average pro Average (B) les generated by alternative low-cell alternative by generated les 'n 'n 004000 2000 3'End 5'End " le of H3K4me3 over all over H3K4me3 of le 1000 cellChIPShenmerged 5000 cellChIPBrind’Amour 10000 cellChIPBrind’Amour 100000 cellChIPBrind’Amour 100 cellChIPmerged 500 cellChIPmerged 1000 cellChIPmerged 3000 cellChIPmerged 2 millioncellbenchChIP HK4me3 ChIP-seq ii platform uidic Supplemental Figure 10

A H3K4me1 H3K27ac H3K36me3

1.00 0.94 0.76 0.82 2M 1.00 0.88 0.65 0.69 2M 1.00 0.83 0.88 2M

0.94 1.00 0.83 0.86 2M 0.88 1.00 0.58 0.64 2M

0.83 1.00 0.93 3000

0.76 0.83 1.00 0.95 3000 0.65 0.58 1.00 0.87 3000

0.88 0.93 1.00 3000 0.82 0.86 0.95 1.00 3000 0.69 0.64 0.87 1.00 3000 2M 2M 2M 2M 2M 3000 3000 3000 3000 3000 3000

B H3K4me3 H3K4me1 H3K27ac H3K36me3

3000 cell micro!uidic ChIP 4714 20598 1520 11237 24662 9172 2071 5675 6669 4659 8465 5273 2 million cell bench ChIP

C H3K4me1 H3K27ac H3K36me3 3000 cells 3000 cells 3000 cells

Replicate 1 5129 35899 8318 4522 7746 2712 3287 13124 4806 Replicate 2

Supplemental Figure 10. PnP-ChIP-seq using small cell quantities using MNase shearing on 15,000 mESCs for H3K4me1, H3K27ac and H3K36me3. (A) Cross-correlations of PnP-ChIP-seq using tag counts of merged peak set for H3K4me1, H3K27ac or H3K36me3 (B) Overlap between de novo peak calls of PnP-ChIP-seq and bulk ChIP-seq. (C) Overlap between de novo peak calls of PnP-ChIP-seq of replicate experiments using 3000 mESC chromatin equivalent as input. Supplemental Figure 11

A Gata6 Lhx1 Hoxb cluster

Scale 1 Mb mm9 Chr11: 96,000,000 96,500,000 97,000,000 Phospho1 Gip Hoxb9 Snx11 Sp2 Mrpl45 Abi3 Snf8 Hoxb8 Cbx1 Sp6 Gpr179 Gngt2 Ube2z Hoxb7 Nfe2l1 Tbx21 Socs7 B4galnt2 Ttll6 Skap1 Pnpo Kpnb1 Igf2bp1 Hoxb6 Copz2 Tbkbp1 Arhgap23 Atp5g1 Hoxb5 Cdk5rap3 Npepps Calcoco2 Hoxb4 Prr15l E130012A19Rik Hoxb13 Scrn2 Scale 50 kb mm9 Scale 50 kb mm9 AK078566 Lrrc46 Chr18: 11,050,000 11,100,000 Chr11: 11,150,000 84,350,000 11,200,000 Hoxb384,400,000 Mrpl10 AK033147 Aatf Hoxb2 Osbpl7 Gata6 Lhx1 Hoxb1 1.5 - 1.5 - 2.5 - Reference 0 _ 0 _ 0 _ 1.5 - 1.5 - 2.5 - 3000 (MNase) 0 _ 0 _ 0 _ 1.5 - 1.5 - 2.5 - Zhang et al. (1000) 0 _ 0 _ 0 _ 1.5 - 1.5 - 2.5 - 3000 (sonication)

H3K27me3 0 _ 0 _ 0 _ 1.5 - 1.5 - 2.5 - 1000 (sonication) 0 _ 0 _ 0 _

H3K27me3 ChIP-seq B 2 million cell bench ChIP 3000 cells bulk sonication 1000 cells bulk sonication 3000 cells of 15k MNase 1000 cells ChIP Zhang Read count Per Million mapped reads Read count Per 0.08 0.10 0.12 0.14

0 25000 Scale 200 kb mm9 C Chr4: 3,500,000 Vmn1r2 Tmem68 Lyn Sdr16c5 Vmn1r3 Tgs1 Rps20 Mos Plag1 Chchd7 17 - Reference 2 _ Bulk-Br4-ENC_H3K9me3_rep2_SRR566936_R1 17 - 3000 H3K9me3 2 _

Scale 2 Mb mm9 Chr4: 144,000,000 144,500,000 145,000,000 145,500,000 146,000,000 146,500,000 147,000,000 147,500,000 148,000,000 Gm13023 Pramel4 Gm13128 Gm13177 Dhrs3 Tnfrsf8 Gm13051 Gm13138 Zfp600 Gm13139 Gm13154 Zfp933 Mad2l2 Ubiad1 Srm Casz1 Gm13084 Gm13102 Gm13125 Gm436 Vps13d Gm13242 Gm13251 Gm13051 Gm13138 2610305D13Rik Miip Fbxo6 Mtor Pex14 Gm13103 Oog3 Aadacl3 Gm438 Tnfrsf1b Cdv3 Gm13235 Zfp534 Fv1 Fbxo44 Angptl7 Dffa Pramef6 Oog2 Gm13124 Vmn2r-ps14 Gm13154 Gm13157 Nppb Ptchd2 Exosc10 Gm13109 C87977 Gm13178 RP23-282C23.3 Mfn2 Fbxo2 Masp2 Gm13101 Pramel5 Plod1 Tardbp Pramef17 Gm13119 2510039O18Rik Gm572 Pramef12 Nppa 1700012P22Rik Clcn6 9430007A20Rik Mthfr Agtrap 2610109H07Rik Reference17 - 2 _ 17 -

3000 2 _ H3K9me3

D % of tags in major Satellite % of tags in SineB (Alu) repeat 10 0.02

8 0.015

6 0.01 4 % recovery % recovery 0.005 2

0 0

Supplemental Figure 11. PnP-ChIP-seq of H3K27me3 and H3K9me3. (A) Examplary genome browser views for PnP-ChIP-seq of H3K27me3. (B) Average pro"le of H3K27me3 over all H3K27me3 peaks in mESCs (van Mierlo et al., 2019) of pro"les generated by PnP-ChIP-seq as compared to bulk ChIP-seq pro"les and an alternative low- cell input ChIP-seq method called STAR ChIP-seq (Zhang et al., 2016). The start and end of the peaks are indicated with 5’end and 3’end, respectively. (C) Genome browser views for PnP-ChIP-seq of H3K9me3. The bottom example shows a genomic region which is highly enriched for major satellite repeats. (D) Presence of PnP-ChIP-seq sequence tags of various hPTMs in two types of repeats, con"rming the presence of H3K9me3 over major satellites. Supplemental Figure 12

H3K4me3 H3K27ac H3K4me1 8 8 8

6 6 6

4 4 4 2i (log $ normalized counts) 2i (log $ normalized counts) 2i (log $ normalized counts)

2 2 2 2 4 6 8 2 4 6 8 2 4 6 8 serum (log$ normalized counts) serum (log$ normalized counts) serum (log$ normalized counts) Supplemental Figure 12. PnP-ChIP-seq allows to detect signi"cant di#erences in the closely-related cell types 2i and serum mESCs. Scatterplot of DESeq2-normalised tag counts (replicate means) of H3K4me3, H3K27ac and H3K4me1 PnP-ChIP-seq on 2i and serum mESCs. Signi"cant enrichment (FDR-adjusted p-value < 0.05) depicted in red. We detect in total 25,617 H3K4me3 peaks (3,459 peaks signi"cantly higher in either 2i ESCs or serum ESCs), 8,329 H3K27ac peaks (161 peaks signi"cantly higher in either 2i ESCs or serum ESCs), and 5,752 H3K4me1 peaks (41 peaks signi"cantly higher in either 2i mESCs or serum mESCs). Please "nd more details in Supplemental Table S1. Supplemental Figure 13

A H3K4me3 ChIP-seq, signi"cantly higher in 2i mESCs Tex14 Ubc Tle6

Scale 50 kb mm9 Scale 1 kb mm9 Scale 5 kb mm9 Chr11: 87,300,000 87,350,000 Chr5: 125,870,000 Chr10: 81,060,000 H3K4me3peaks_all H3K4me3peaks_sig 8 - 10 - 8 - H3K4me3_2i_r1 0 _ 0 _ 0 _ 8 - 10 - 8 - H3K4me3_2i_r2 0 _ 0 _ 0 _ 8 - 10 - 8 - H3K4me3_FCS_r1 0 _ 0 _ 0 _ 8 - 10 - 8 - H3K4me3_FCS_r2 0 _ 0 _ 0 _ 8 - 10 - 8 - H3K4me3_FCS_r3 0 _ 0 _ 0 _ Tex14 Ubc Tle2 Tle6

B H3K4me3 ChIP-seq, signi"cantly higher in serum mESCs Lin28b Dnmt3l Cdk12 Scale 50 kb mm9 Chr10: 45,150,000 45,200,000 Scale 10 kb mm9 Scale 20 kb mm9 H3K4me3peaks_all Chr10: 77,520,000 Chr11: 98,100,000 H3K4me3peaks_sig 6 - H3K4me3_2i_r1 10 - 12 - 0 _ 6 - 0 _ 0 _ H3K4me3_2i_r2 10 - 12 - 0 _ 6 - 0 _ 0 _ H3K4me3_FCS_r1 10 - 12 - 0 _ 6 - 0 _ 0 _ H3K4me3_FCS_r2 10 - 12 - 0 _ 6 - 0 _ 0 _ H3K4me3_FCS_r3 10 - 12 - 0 _ Lin28b 0 _ 0 _ Lin28b Dnmt3l Cdk12

Supplemental Figure 13. PnP-ChIP-seq allows the detection of signi"cant di #erences in H3K4me3 between the closely-related cell types 2i and serum ESCs (serum is abbreviated as FCS (foetal calf serum)). (A) Genome browser views for H3K4me3 PnP-ChIP-seq of 2i and serum mESCs for 3 genes that are known to be higher expressed in 2i mESCs. (B) Genome browser views for H3K4me3 PnP-ChIP-seq of 2i and serum mESCs for 3 genes that are known to be higher expressed in serum mESCs. r = replicate; H3K4me3peaks_all = merge track of all H3K4me3 peaks detected in 2i mESCs and serum mESCs. H3K4me3peaks_sig = merge track of all H3K4me3 peaks that are signi"cantly increased in either 2i mESCs or serum mESCs. Signi"cant di#erences in gene expression between 2i and serum ESCs from Marks et al., 2012 (PMID: 22541430). Supplemental Figure 14

A H3K27ac ChIP-seq, signi"cantly higher in 2i mESCs Fam166a Fads2 Sccpdh

Scale 2 kb mm9 Scale 20 kb mm9 Scale 20 kb mm9 Chr2: 25,075,000 Chr19: 10,200,000 Chr1: 181,600,000 H3K4me3peaks_all H3K4me3peaks_sig 8 - 8 - 3 - H3K4me3_2i_r1 0 _ 0 _ 0 _ 8 - 8 - 3 - H3K4me3_2i_r2 0 _ 0 _ 0 _ 8 - 8 - 3 - H3K4me3_FCS_r1 0 _ 0 _ 0 _ 8 - 8 - 3 - H3K4me3_FCS_r2 0 _ 0 _ 0 _ 8 - 8 - 3 - H3K4me3_FCS_r3 0 _ 0 _ 0 _ Fam166a Fads2 Sccpdh Tubb2c H3K27acpeaks_all H3K27acpeaks_sig 6 - 2 - 1 - H3K27ac_2i_r1 0 _ 0 _ 0 _ 6 - 2 - 1 - H3K27ac_2i_r2 0 _ 0 _ 0 _ 6 - 2 - 1 - H3K27ac_FCS_r1 0 _ 0 _ 0 _ 6 - 2 - 1 - H3K27ac_FCS_r2 0 _ 0 _ 0 _ H3K4me1peaks_all H3K4me1peaks_sig 3 - 3 - 2 - H3K4me1_2i_r1 0 _ 0 _ 0 _ 3 - 3 - 2 - H3K4me1_2i_r2 0 _ 0 _ 0 _ 3 - 3 - 2 - H3K4me1_FCS_r1 0 _ 0 _ 0 _ 3 - 3 - 2 - H3K4me1_FCS_r2 0 _ 0 _ 0 _

B H3K27ac ChIP-seq, signi"cantly higher in serum mESCs

Xkr6 Ripk2 Ttc32

Scale 100 kb mm9 Scale 100 kb mm9 Scale 10 kb mm9 Chr14: 64,300,000 64,400,000 Chr4: 16,200,000 16,300,000 16,400,000 Chr12: H3K4me3peaks_all H3K4me3peaks_sig 6 - 4 - 6 - H3K4me3_2i_r1 0 _ 0 _ 0 _ 6 - 4 - 6 - H3K4me3_2i_r2 0 _ 0 _ 0 _ 6 - 4 - 6 - H3K4me3_FCS_r1 0 _ 0 _ 0 _ 6 - 4 - 6 - H3K4me3_FCS_r2 0 _ 0 _ 0 _ 6 - 4 - 6 - H3K4me3_FCS_r3 0 _ 0 _ 0 _ Xkr6 Ripk2 Wdr35 Ttc32 H3K27acpeaks_all H3K27acpeaks_sig 2 - 2 - 4 - H3K27ac_2i_r1 0 _ 0 _ 0 _ 2 - 2 - 4 - H3K27ac_2i_r2 0 _ 0 _ 0 _ 2 - 2 - 4 - H3K27ac_FCS_r1 0 _ 0 _ 0 _ 2 - 2 - 4 - H3K27ac_FCS_r2 0 _ 0 _ 0 _ H3K4me1peaks_all H3K4me1peaks_sig 2 - 2 - 2 - H3K4me1_2i_r1 0 _ 0 _ 0 _ 2 - 2 - 2 - H3K4me1_2i_r2 0 _ 0 _ 0 _ 2 - 2 - 2 - H3K4me1_FCS_r1 0 _ 0 _ 0 _ 2 - 2 - 2 - H3K4me1_FCS_r2 0 _ 0 _ 0 _ Supplemental Figure 14. PnP-ChIP-seq allows the detection of signi"cant di#erences in H3K27ac between the closely-related cell types 2i and serum ESCs (serum is abbreviated as FCS (foetal calf serum)). (A) Genome browser views for PnP-ChIP-seq of loci showing a signi"cant increase in H3K27ac in 2i mESCs as compared to serum mESCs (boxed) (B) Genome browser views for PnP-ChIP-seq of loci showing a signi"cant increase in H3K27ac in serum mESCs as compared to 2i mESCs (boxed). r = replicate; H3*peaks_all = merge track of all peaks of a hPTM detected in 2i mESCs and serum mESCs. H3*peaks_sig = merge track of all peaks of a hPTM that are signi"cantly increased in either 2i mESCs or serum mESCs. Supplemental Figure 15

A H3K4me1 ChIP-seq, signi"cantly higher in 2i mESCs Oas1g Ppp1r2 Ppara

Scale 10 kb mm9 Scale 10 kb mm9 Scale 20 kb mm9 Chr5: 121,350,000 Chr16: 31,260,000 31,270,000 Chr15: 85,600,000 H3K4me3peaks_all H3K4me3peaks_sig 3 - 2 - 4 - H3K4me3_2i_r1 0 _ 0 _ 0 _ 3 - 2 - 4 - H3K4me3_2i_r2 0 _ 0 _ 0 _ 3 - 2 - 4 - H3K4me3_FCS_r1 0 _ 0 _ 0 _ 3 - 2 - 4 - H3K4me3_FCS_r2 0 _ 0 _ 0 _ 3 - 2 - 4 - H3K4me3_FCS_r3 0 _ 0 _ 0 _ Oas1h Oas1a Ppp1r2 Ppara Oas1g 2210021J22Rik H3K27acpeaks_all H3K27acpeaks_sig 2 - 2 - 2 - H3K27ac_2i_r1 0 _ 0 _ 0 _ 2 - 2 - 2 - H3K27ac_2i_r2 0 _ 0 _ 0 _ 2 - 2 - 2 - H3K27ac_FCS_r1 0 _ 0 _ 0 _ 2 - 2 - 2 - H3K27ac_FCS_r2 0 _ 0 _ 0 _ H3K4me1peaks_all H3K4me1peaks_sig 2 - 2 - 1 - H3K4me1_2i_r1 0 _ 0 _ 0 _ 2 - 2 - 1 - H3K4me1_2i_r2 0 _ 0 _ 0 _ 2 - 2 - 1 - H3K4me1_FCS_r1 0 _ 0 _ 0 _ 2 - 2 - 1 - H3K4me1_FCS_r2 0 _ 0 _ 0 _

B H3K4me1 ChIP-seq, signi"cantly higher in serum mESCs Cdkn1c Arrdc3 Hist1h1e

Scale 5 kb mm9 Scale 10 kb mm9 Scale 5 kb mm9 Chr7: 150,645,000 150,650,000 Chr13: 81,030,000 81,040,000 81,050,000 Chr13: 23,710,000 23,715,000 23,720,000 H3K4me3peaks_all H3K4me3peaks_sig 4 - 12 - 8 - H3K4me3_2i_r1 0 _ 0 _ 0 _ 4 - 12 - 8 - H3K4me3_2i_r2 0 _ 0 _ 0 _ 4 - 12 - 8 - H3K4me3_FCS_r1 0 _ 0 _ 0 _ 4 - 12 - 8 - H3K4me3_FCS_r2 0 _ 0 _ 0 _ 4 - 12 - 8 - H3K4me3_FCS_r3 0 _ 0 _ 0 _ Cdkn1c Arrdc3 Hist1h2be Hist1h1e H3K27acpeaks_all H3K27acpeaks_sig 2 - 4 - 6 - H3K27ac_2i_r1 0 _ 0 _ 0 _ 2 - 4 - 6 - H3K27ac_2i_r2 0 _ 0 _ 0 _ 2 - 4 - 6 - H3K27ac_FCS_r1 0 _ 0 _ 0 _ 2 - 4 - 6 - H3K27ac_FCS_r2 0 _ 0 _ 0 _ H3K4me1peaks_all H3K4me1peaks_sig 4 - 2 - 5 - H3K4me1_2i_r1 0 _ 0 _ 0 _ 4 - 2 - 5 - H3K4me1_2i_r2 0 _ 0 _ 0 _ 4 - 2 - 5 - H3K4me1_FCS_r1 0 _ 0 _ 0 _ 4 - 2 - 5 - H3K4me1_FCS_r2 0 _ 0 _ 0 _

Supplemental Figure 15. PnP-ChIP-seq allows the detection of signi"cant di#erences in H3K4me1 between the closely-related cell types 2i and serum ESCs (serum is abbreviated as FCS (foetal calf serum)). (A) Genome browser views for PnP-ChIP-seq of loci showing a signi"cant increase in H3K4me1 in 2i mESCs as compared to serum mESCs (boxed) (B) Genome browser views for PnP-ChIP-seq of loci showing a signi"cant increase in H3K4me1 in serum mESCs as compared to 2i mESCs (boxed). r = replicate; H3*peaks_all = merge track of all peaks of a hPTM detected in 2i mESCs and serum mESCs. H3*peaks_sig = merge track of all peaks of a hPTM that are signi"cantly increased in either 2i mESCs or serum mESCs.

Supplemental Figure 16

(x1000)

FSC-A

FSC-A FSC-A

Fluorescent intensity (Hhex::Venus) Fluorescent intensity (MERVL::tdTomato) Fluorescent intensity (Zscan4c::emGFP)

Supplemental Figure 16. Gating of embryonic stem cell subpopulations on !uorescent markers Venus (driven by the Hhex promoter), tdTomato (driven by the MERVL promoter) and emGFP (driven by the Zscan4c promoter). Supplemental Figure 17

R = 0.88 R = 0.89 R = 0.8 15 15 8

K 6 10 10 ov ov

5 5 k 2 k k Normaliz

0 0 0 Normaliz Normaliz 0 5 10 15 0 5 10 15 0 2 4 6 8 Normaliz Normaliz Normaliz

B o

p < 2.2e-16 p = 0.001075 p = 3.962e-14 50 100 150 50 100 150 Normaliz Normaliz 0 0

H H H H R R R R 50 100 150 Normaliz 0

H H

R R Supplemental Figure 17. Epigenome analysis of 2C-like cells as compared to wildtype mESCs. (A) Scatterplot of DESeq2-normalised ChIP-seq tag counts of H3K27ac, H3K4me1 (both from Zhang et al., 2019) and H3K27me3 (Hayashi et al., 2016) of 2C-like cells as compared to wildtype mESCs. DESeq2 analysis detected no signi"cant di#erences (FDR-adjusted p-value < 0.05) in H3K27ac, H3K4me1 or H3K27me3 between 2C-like cells and wildtype mESCs. The axes of the plots are in log%. (B) Boxplots of DESeq2-normalised tag counts of H3K4me3 PnP-ChIP-seq pro"les of Hhex-, MERVL- and Zscan4c-positive ("pos") and negative ("neg") mESC cell populations of all genes, and genes that have been reported to be higher of lower in 2C-like cells as compared to wildtype mESCs (Fu et al., 2019). Signi"cant di#erences between boxplots calculated using the Wilcoxon rank- sum test are indicated.