Library assembly in rounds

Protocol 6 of 7. Comes after whichever of the two ways to read every well back you did. Next is Final cargo ligation.

Join 3 part lists into 13,824 distinct constructs in 3 rounds, each round opening the library with BbsI and ligating one part list into it, then read the finished library back.

Method: iGGA
Part lists: N 24, DBD 24, C 24
Parts: 72 blocks, assembled from 131 oligos
Constructs: 13,824 distinct
Rounds: 3, one a part list
Opened with: BbsI-HF (R3539)
Released with: BsaI-HFv2 (R3733)
Blunt enzymes: SrfI and PmeI
Entry overhangs: AGGA, AGAT and GCAT
Cloning scar: TTCC
Barcode: 11 bp, block 41 bp
Codon usage: human
Product: AP-1 DESynR round 3, 6435 bp
Completeness: 0.99 chance nothing is missing, 195,386 colonies at the end
Amino acids changed: 27 over 26 part ends
Designs read back: 72, every one

Have in hand

part carrier plate one pCR-Blunt II-TOPO a part, one part a well, selected on KanR
clonal picked plate one well a picked colony, each well one clone
well calls a pass or a fail a well, and which design each well holds
block vector 1 DMX-iGGA N, ready to open

Leaves you with

library prep, round 3 the pooled library after round 3, as a plasmid prep

1 number in this protocol has no source.

Materials

Name	Supplier	Catalogue	Storage	Per run	Note	Source
N part list			-20 °C	24 members, pooled	assembled from the oligo pool; pool.tsv says which oligos
DBD part list			-20 °C	24 members, pooled	assembled from the oligo pool; pool.tsv says which oligos
C part list			-20 °C	24 members, pooled	assembled from the oligo pool; pool.tsv says which oligos
DMX-iGGA vector			-20 °C	3, one a position	opened by the internal digest in round 1, and to take each block's cargo. A part enters on its own position's entry overhang, so there is one a position: DMX-iGGA N (block-vector-1.dna), DMX-iGGA DBD (block-vector-2.dna) and DMX-iGGA C (block-vector-3.dna)
BbsI-HF	New England Biolabs	R3539	-20 °C	2.5 µL per digest	opens the library, excising its stuffer
BsaI-HFv2	New England Biolabs	R3733	-20 °C	2.5 µL per digest	releases a part from its own block
SrfI	New England Biolabs	R0629	-20 °C	2.5 µL per digest	cuts the excised stuffer, so it cannot ligate back
PmeI	New England Biolabs	R0560	-20 °C	2.5 µL per digest	cuts the donor backbone, so it cannot ligate back
CutSmart Buffer			-20 °C		both digests run in it
T7 DNA Ligase	NEB	#M0318L	-20 °C		3,000,000 U/mL. Refuses blunt ends, which is what removes escapees.	^[1]
StickTogether DNA Ligase Buffer	NEB	#B0535S	-20 °C		Supplied 2x. 1x is 7.5% PEG 6000, well under the 20% that would force blunt ligation.	^[1]
SPRI paramagnetic beads				2 volumes after a digest, 1 after a ligation	the two ratios differ; both elute in water
Endura ElectroCompetent Cells	Lucigen	60242-2	-80 °C	4 aliquots, one a pulse, 25 µL each	pulsed at 1800 V, 600 Ω and 10 µF in a 1 mm cuvette	^[2]
Recovery medium				one outgrowth per round
Selective broth and plates					LB with 50 µg/mL kanamycin, the destination vector's own marker (KanR)
Plasmid prep kit				one prep per round
Electroporation cuvettes				4, one a pulse

Equipment: Incubator or heat block at 37 °C, Magnetic rack for the bead clean-ups, Electroporator and cuvettes, Shaking incubator at 30 °C, Spectrophotometer (NanoDrop) or fluorometer (Qubit), Long-read sequencer, for the linkage read of the finished library, Short-read sequencer, for the representation read of the finished library

T7 DNA Ligase: never PEG Never add PEG to a T7 ligase reaction. Blunt-end ligation is forced above 20% (w/v) PEG 6000, and StickTogether buffer already carries 7.5%; the refusal that removes escapees holds only while nothing adds more.^[1]
T7 DNA Ligase: never heat inactivation Do not heat-inactivate T7 DNA ligase in PEG: transformation will be inhibited. Go from the ligation straight to the clean-up.^[1]

Caution: Keep the cells and the cuvette on ice; a warm cuvette arcs.

Pool the part lists

1Pool each part list

Clean up each assembly and measure each concentration.
Pool the members of each part list in equal picomoles, one tube a position: each member gives the position's total divided by its member count, which is unequal masses because the members differ in length.
Bring each pool to at least 22.3 ng/µL. The round's 50 µL digest leaves 45 µL for the DNA after its two 2.5 µL enzymes, and the whole total has to arrive in it.

Expected result

N: one tube, 24 members, at least 1,000 ng at 22.3 ng/µL or above, 0.11 pmol of each member.
DBD: one tube, 24 members, at least 1,000 ng at 22.3 ng/µL or above, 0.219 pmol of each member.
C: one tube, 24 members, at least 1,000 ng at 22.3 ng/µL or above, 0.193 pmol of each member.

Troubleshooting

One member is much more dilute than the rest: Pool to the lowest member rather than to the mean; a member short here is short in every round after it.
The pool is below 22.3 ng/µL: Concentrate it, by SPRI at the round's own 2x ratio, or scale the digest up so the same mass arrives in a larger volume.

Notes

Library coverage is counted on equally represented members, so an uneven pool loses members that no later round can put back.
Equal picomoles are unequal masses: a short member weighs less than a long one for the same number of molecules, and weighing them equally would not pool them equally.^[3]
Every round's digest table is laid out at 22.3 ng/µL, which is the floor and not a target: at it the pool takes almost the whole tube and the buffer and water line all but disappears. Concentrate well above it and pipette less pool; the buffer and water line makes the volume up either way.
N: 0.11 pmol a member is 9.99 ng of its shortest at 147 bp and 78.1 ng of its longest at 1,149 bp.
DBD: 0.219 pmol a member is 39.2 ng of its shortest at 291 bp and 48.9 ng of its longest at 363 bp.
C: 0.193 pmol a member is 15.8 ng of its shortest at 133 bp and 125 ng of its longest at 1,048 bp.

Round 1

2Round 1: open DMX-iGGA with BbsI

Set the digest below up with BbsI alone, and start the program.
After the first hour add SrfI and run the second.

Digest with BbsI and SrfI

Reactions mix includes 10% extra

Component	Stock	Final	1 rxn (µL)	Mix for 1 (µL)
DMX-iGGA^[3]		0.301 pmol (1000 ng)	1	each tube
BbsI-HF (R3539)^[4]	R3539L, 20 U/µL	50 units	2.5	2.75
SrfI (R0629)^[4]	R0629L, 20 U/µL	50 units	2.5	2.75
CutSmart Buffer and nuclease-free water		to 50 µL	44	48.4
Total			50	53.9 mix only

Put 49 µL of mix in each tube, then add 1 µL DMX-iGGA.

Digest with BbsI and SrfI · 2 h

Step	Temperature	Time	Cycles
BbsI	37 °C	1 h	1^[5]
BbsI and SrfI	37 °C	1 h	1^[5]

Digest with BbsI and SrfI

Expected result

DMX-iGGA opens on AGGA and TTCC, leaving 5363 bp of backbone.
The 30 bp internal stuffer comes out, and SrfI cuts it so it cannot go back in.

Troubleshooting

Mostly empty library later: The destination was not opened to completion. Run a little of the digest on a gel before ligating the next round.

Notes

BbsI is the enzyme the product keeps sites for. That is the design working, not a site left behind.

3Round 1: release the N part list with BsaI

Set the digest below up with BsaI alone, and start the program.
After the first hour add PmeI and run the second.

Digest with BsaI and PmeI

Reactions mix includes 10% extra

Component	Stock	Final	1 rxn (µL)	Mix for 1 (µL)
N part list^[3]		2.65 pmol (1000 ng)	44.8	each tube
BsaI-HFv2 (R3733)^[4]	R3733L, 20 U/µL	50 units	2.5	2.75
PmeI (R0560)^[4]	R0560L, 10 U/µL	25 units	2.5	2.75
CutSmart Buffer and nuclease-free water		to 50 µL	0.16	0.176
Total			50	5.68 mix only

Put 5.16 µL of mix in each tube, then add 44.8 µL N part list.

Digest with BsaI and PmeI · 2 h

Step	Temperature	Time	Cycles
BsaI	37 °C	1 h	1^[5]
BsaI and PmeI	37 °C	1 h	1^[5]

Digest with BsaI and PmeI

Expected result

Every member of the N part list is released, about 559 bp, on the same two overhangs the destination now offers.
PmeI cuts the two external stuffers left behind, so neither can ligate back.

Troubleshooting

A member is missing from the library later: Its block was under-represented in the pool, or its own digest was incomplete. Check the pool before repeating the round.

Notes

One tube takes the whole part list: every member carries the same stuffers and differs only in its coding bases and its barcode.^[6]
The amounts are weighed at the pool's mean block length, 613 bp.

4Round 1: clean both digests up

Add 2 volumes of SPRI paramagnetic beads to each digest and elute in water.
Measure both concentrations; the ligation table asks for picomoles, not nanograms.

Expected result

DMX-iGGA, opened: 0.00605 pmol is 20 ng at 5363 bp.
N part list, released: 0.00605 pmol is 2.08 ng at 559 bp.

Troubleshooting

Too dilute for the ligation: Elute in a smaller volume, or scale the ligation up.

Notes

2 volumes here and 1 after the ligation. The two ratios differ; they are not one number used twice.

5Round 1: ligate the N part list into the library

T7 DNA Ligase: never PEG Never add PEG to a T7 ligase reaction. Blunt-end ligation is forced above 20% (w/v) PEG 6000, and StickTogether buffer already carries 7.5%; the refusal that removes escapees holds only while nothing adds more.^[1]

Pipette the two DNAs into the tube first, then the ligase, its buffer and water.
Hold at room temperature for 30 minutes.

Ligation

Reactions mix includes 10% extra

Component	Final	1 rxn (µL)	Mix for 1 (µL)
DMX-iGGA, opened^[3]	0.00605 pmol (20 ng)	1	each tube
N part list, released^[3]	0.00605 pmol (2.08 ng)	1	each tube
T7 DNA Ligase in StickTogether DNA Ligase Buffer, and nuclease-free water	to 200 µL	198	218
Total		200	218 mix only

Put 198 µL of mix in each tube, then add 1 µL DMX-iGGA, opened, 1 µL N part list, released.

Ligation

DMX-iGGA N

Explore the mapClose the map

Open it in a new tab

AP-1 DESynR round 1

Explore the mapClose the map

Open it in a new tab

Round 1 of 3 joins N_JUN at position N, entering on AGGA and leaving TTCC: what it opens above, what it makes below^[7]

Expected result

Nothing visible. 0.00605 pmol of the released part list against 0.00605 pmol of the opened library is the 1:1 molar ratio.

Troubleshooting

Few colonies after this round: Check that both digests went to completion; an uncut end cannot ligate.

Notes

The 1:1 ratio is the method's, and it is a molar ratio: the shorter fragment weighs less at the same one.^[5]
The method names T7 DNA Ligase and StickTogether DNA Ligase Buffer and publishes neither the units nor the volume, so the last line of the table is the supplier's own.

6Round 1: clean the ligation up

Add 1 volume of SPRI paramagnetic beads and elute in water.
Measure the concentration.

Expected result

Enough for one electroporation: 100 ng is 0.0264 pmol at 6161 bp.

Troubleshooting

The eluate is still salty: Repeat the clean-up; salt carried into the cuvette arcs the pulse.

Notes

One volume here, not the 2 the digests took. Eluting in water rather than a buffer is what keeps the salt out of the cuvette.

7Round 1: electroporate into Endura ElectroCompetent Cells

Caution: Keep the cells and the cuvette on ice; a warm cuvette arcs.

Thaw one aliquot of Endura ElectroCompetent Cells on ice, 25 µL a pulse.
Add at most 100 ng of the purified ligation and mix without making bubbles.
Pulse at 1800 V, 600 Ω and 10 µF in a 1 mm cuvette.

Expected result

A pulse with no arc, and a time constant of 3.5 to 4.5 ms.

Troubleshooting

The cuvette arcs: The DNA carries salt. Clean it up again, elute in water, and use a fresh aliquot of cells.

Notes

The method names the instrument and none of the settings, so the program is the cells' own: it is keyed by 60242-2 and changes when the cells do. The materials table says where it was read.

8Round 1: recover and grow at 30 °C

Add recovery medium straight away and shake for the first hour.
Plate a measured dilution of the recovery on 50 µg/mL kanamycin as round 1 titre, and grow the rest in selective broth.
Plate the no-donor ligation from the same digest as round 1 no-donor control, at the same dilution.

Recovery and outgrowth at 30 °C · 13 h

Step	Temperature	Time	Cycles
Recovery, shaking	30 °C	1 h	1^[5]
Outgrowth	30 °C	12 h	1^[5]

Recovery and outgrowth at 30 °C

Expected result

At least 183 net colonies, scaled up from the dilution: the floor for the 0.99 chance this design asked for that none of its 24 distinct products is missing, equally represented.
At that count the chance a named product is missing is 4.15 × 10⁻⁴.
The round 1 no-donor control should be near empty beside it; its colonies come off the count.

Counted on round 1 titre less round 1 no-donor control Dilution factor

— net colonies.

Troubleshooting

Fewer net colonies than the count above: The round has lost library members and no later round can put them back. Electroporate more of the ligation, or run the round again.
The no-donor control is not empty: The destination was not opened to completion, or the blunt chopper missed. What grows is the round before's library, one position short, and it reaches the linkage read as truncated members.

Notes

Both steps run at 30 °C and not at 37 °C. That is a library precaution rather than an oversight.
The program carries the shorter outgrowth; 12 to 16 hours is the range the method gives.
The 0.99 was chosen for this run, and follows from the representation the screen downstream asks for. No source sets it, and nothing here defaults it; it works out at 8x this round's products.
The colony floor is what the chance of a missing member comes to over a library whose members are equally represented.^[8]
The control measures the chain the design rests on — two cuts, a blunt chopper, a ligase that refuses blunt ends and the 2x clean-up — rather than assuming it.

9Round 1: prep the library

Harvest the whole culture rather than a single colony.
Prep the plasmid and measure the concentration.

Expected result

One pool of plasmid, 6161 bp per molecule. This is the destination round 2 opens.
BbsI cuts it in 2 places, around the stuffer the part carried in. BsaI cuts it in 2 places, as it cut the destination: the part brought none in.

Troubleshooting

A low yield: The culture was grown too briefly, or the antibiotic was wrong for this backbone. Check the vector's own marker.

Notes

Harvesting the whole culture is what keeps the library a library; picking colonies here throws away everything not picked.

Round 2

10Round 2: open AP-1 DESynR round 1 with BbsI

Set the digest below up with BbsI alone, and start the program.
After the first hour add SrfI and run the second.

Digest with BbsI and SrfI

Reactions mix includes 10% extra

Component	Stock	Final	1 rxn (µL)	Mix for 1 (µL)
AP-1 DESynR round 1^[3]		0.264 pmol (1000 ng)	1	each tube
BbsI-HF (R3539)^[4]	R3539L, 20 U/µL	50 units	2.5	2.75
SrfI (R0629)^[4]	R0629L, 20 U/µL	50 units	2.5	2.75
CutSmart Buffer and nuclease-free water		to 50 µL	44	48.4
Total			50	53.9 mix only

Put 49 µL of mix in each tube, then add 1 µL AP-1 DESynR round 1.

Digest with BbsI and SrfI · 2 h

Step	Temperature	Time	Cycles
BbsI	37 °C	1 h	1^[5]
BbsI and SrfI	37 °C	1 h	1^[5]

Digest with BbsI and SrfI

Expected result

AP-1 DESynR round 1 opens on AGAT and TTCC, leaving 6131 bp of backbone.
The 30 bp internal stuffer comes out, and SrfI cuts it so it cannot go back in.

Troubleshooting

Mostly empty library later: The destination was not opened to completion. Run a little of the digest on a gel before ligating the next round.

Notes

BbsI is the enzyme the product keeps sites for. That is the design working, not a site left behind.

11Round 2: release the DBD part list with BsaI

Set the digest below up with BsaI alone, and start the program.
After the first hour add PmeI and run the second.

Digest with BsaI and PmeI

Reactions mix includes 10% extra

Component	Stock	Final	1 rxn (µL)	Mix for 1 (µL)
DBD part list^[3]		5.25 pmol (1000 ng)	44.8	each tube
BsaI-HFv2 (R3733)^[4]	R3733L, 20 U/µL	50 units	2.5	2.75
PmeI (R0560)^[4]	R0560L, 10 U/µL	25 units	2.5	2.75
CutSmart Buffer and nuclease-free water		to 50 µL	0.16	0.176
Total			50	5.68 mix only

Put 5.16 µL of mix in each tube, then add 44.8 µL DBD part list.

Digest with BsaI and PmeI · 2 h

Step	Temperature	Time	Cycles
BsaI	37 °C	1 h	1^[5]
BsaI and PmeI	37 °C	1 h	1^[5]

Digest with BsaI and PmeI

Expected result

Every member of the DBD part list is released, about 255 bp, on the same two overhangs the destination now offers.
PmeI cuts the two external stuffers left behind, so neither can ligate back.

Troubleshooting

A member is missing from the library later: Its block was under-represented in the pool, or its own digest was incomplete. Check the pool before repeating the round.

Notes

One tube takes the whole part list: every member carries the same stuffers and differs only in its coding bases and its barcode.^[6]
The amounts are weighed at the pool's mean block length, 309 bp.

12Round 2: clean both digests up

Add 2 volumes of SPRI paramagnetic beads to each digest and elute in water.
Measure both concentrations; the ligation table asks for picomoles, not nanograms.

Expected result

AP-1 DESynR round 1, opened: 0.0053 pmol is 20 ng at 6131 bp.
DBD part list, released: 0.0053 pmol is 0.83 ng at 255 bp.

Troubleshooting

Too dilute for the ligation: Elute in a smaller volume, or scale the ligation up.

Notes

2 volumes here and 1 after the ligation. The two ratios differ; they are not one number used twice.

13Round 2: ligate the DBD part list into the library

T7 DNA Ligase: never PEG Never add PEG to a T7 ligase reaction. Blunt-end ligation is forced above 20% (w/v) PEG 6000, and StickTogether buffer already carries 7.5%; the refusal that removes escapees holds only while nothing adds more.^[1]

Pipette the two DNAs into the tube first, then the ligase, its buffer and water.
Hold at room temperature for 30 minutes.

Ligation

Reactions mix includes 10% extra

Component	Final	1 rxn (µL)	Mix for 1 (µL)
AP-1 DESynR round 1, opened^[3]	0.0053 pmol (20 ng)	1	each tube
DBD part list, released^[3]	0.0053 pmol (0.83 ng)	1	each tube
T7 DNA Ligase in StickTogether DNA Ligase Buffer, and nuclease-free water	to 200 µL	198	218
Total		200	218 mix only

Put 198 µL of mix in each tube, then add 1 µL AP-1 DESynR round 1, opened, 1 µL DBD part list, released.

Ligation

AP-1 DESynR round 1

Explore the mapClose the map

Open it in a new tab

AP-1 DESynR round 2

Explore the mapClose the map

Open it in a new tab

Round 2 of 3 joins DBD_JUN at position DBD, entering on AGAT and leaving TTCC: what it opens above, what it makes below^[7]

Expected result

Nothing visible. 0.0053 pmol of the released part list against 0.0053 pmol of the opened library is the 1:1 molar ratio.

Troubleshooting

Few colonies after this round: Check that both digests went to completion; an uncut end cannot ligate.

Notes

The 1:1 ratio is the method's, and it is a molar ratio: the shorter fragment weighs less at the same one.^[5]
The method names T7 DNA Ligase and StickTogether DNA Ligase Buffer and publishes neither the units nor the volume, so the last line of the table is the supplier's own.

14Round 2: clean the ligation up

Add 1 volume of SPRI paramagnetic beads and elute in water.
Measure the concentration.

Expected result

Enough for one electroporation: 100 ng is 0.0255 pmol at 6368 bp.

Troubleshooting

The eluate is still salty: Repeat the clean-up; salt carried into the cuvette arcs the pulse.

Notes

One volume here, not the 2 the digests took. Eluting in water rather than a buffer is what keeps the salt out of the cuvette.

15Round 2: electroporate into Endura ElectroCompetent Cells

Caution: Keep the cells and the cuvette on ice; a warm cuvette arcs.

Thaw one aliquot of Endura ElectroCompetent Cells on ice, 25 µL a pulse.
Add at most 100 ng of the purified ligation and mix without making bubbles.
Pulse at 1800 V, 600 Ω and 10 µF in a 1 mm cuvette.

Expected result

A pulse with no arc, and a time constant of 3.5 to 4.5 ms.

Troubleshooting

The cuvette arcs: The DNA carries salt. Clean it up again, elute in water, and use a fresh aliquot of cells.

Notes

The method names the instrument and none of the settings, so the program is the cells' own: it is keyed by 60242-2 and changes when the cells do. The materials table says where it was read.

16Round 2: recover and grow at 30 °C

Add recovery medium straight away and shake for the first hour.
Plate a measured dilution of the recovery on 50 µg/mL kanamycin as round 2 titre, and grow the rest in selective broth.
Plate the no-donor ligation from the same digest as round 2 no-donor control, at the same dilution.

Recovery and outgrowth at 30 °C · 13 h

Step	Temperature	Time	Cycles
Recovery, shaking	30 °C	1 h	1^[5]
Outgrowth	30 °C	12 h	1^[5]

Recovery and outgrowth at 30 °C

Expected result

At least 6,306 net colonies, scaled up from the dilution: the floor for the 0.99 chance this design asked for that none of its 576 distinct products is missing, equally represented.
At that count the chance a named product is missing is 1.74 × 10⁻⁵.
The round 2 no-donor control should be near empty beside it; its colonies come off the count.

Counted on round 2 titre less round 2 no-donor control Dilution factor

— net colonies.

Troubleshooting

Fewer net colonies than the count above: The round has lost library members and no later round can put them back. Electroporate more of the ligation, or run the round again.
The no-donor control is not empty: The destination was not opened to completion, or the blunt chopper missed. What grows is the round before's library, one position short, and it reaches the linkage read as truncated members.

Notes

Both steps run at 30 °C and not at 37 °C. That is a library precaution rather than an oversight.
The program carries the shorter outgrowth; 12 to 16 hours is the range the method gives.
The 0.99 was chosen for this run, and follows from the representation the screen downstream asks for. No source sets it, and nothing here defaults it; it works out at 11x this round's products.
The colony floor is what the chance of a missing member comes to over a library whose members are equally represented.^[8]
The control measures the chain the design rests on — two cuts, a blunt chopper, a ligase that refuses blunt ends and the 2x clean-up — rather than assuming it.

17Round 2: prep the library

Harvest the whole culture rather than a single colony.
Prep the plasmid and measure the concentration.

Expected result

One pool of plasmid, 6368 bp per molecule. This is the destination round 3 opens.
BbsI cuts it in 2 places, around the stuffer the part carried in. BsaI cuts it in 2 places, as it cut the destination: the part brought none in.

Troubleshooting

A low yield: The culture was grown too briefly, or the antibiotic was wrong for this backbone. Check the vector's own marker.

Notes

Harvesting the whole culture is what keeps the library a library; picking colonies here throws away everything not picked.

Round 3

18Round 3: open AP-1 DESynR round 2 with BbsI

Set the digest below up with BbsI alone, and start the program.
After the first hour add SrfI and run the second.

Digest with BbsI and SrfI

Reactions mix includes 10% extra

Component	Stock	Final	1 rxn (µL)	Mix for 1 (µL)
AP-1 DESynR round 2^[3]		0.255 pmol (1000 ng)	1	each tube
BbsI-HF (R3539)^[4]	R3539L, 20 U/µL	50 units	2.5	2.75
SrfI (R0629)^[4]	R0629L, 20 U/µL	50 units	2.5	2.75
CutSmart Buffer and nuclease-free water		to 50 µL	44	48.4
Total			50	53.9 mix only

Put 49 µL of mix in each tube, then add 1 µL AP-1 DESynR round 2.

Digest with BbsI and SrfI · 2 h

Step	Temperature	Time	Cycles
BbsI	37 °C	1 h	1^[5]
BbsI and SrfI	37 °C	1 h	1^[5]

Digest with BbsI and SrfI

Expected result

AP-1 DESynR round 2 opens on GCAT and TTCC, leaving 6338 bp of backbone.
The 30 bp internal stuffer comes out, and SrfI cuts it so it cannot go back in.

Troubleshooting

Mostly empty library later: The destination was not opened to completion. Run a little of the digest on a gel before ligating the next round.

Notes

BbsI is the enzyme the product keeps sites for. That is the design working, not a site left behind.

19Round 3: release the C part list with BsaI

Set the digest below up with BsaI alone, and start the program.
After the first hour add PmeI and run the second.

Digest with BsaI and PmeI

Reactions mix includes 10% extra

Component	Stock	Final	1 rxn (µL)	Mix for 1 (µL)
C part list^[3]		4.64 pmol (1000 ng)	44.8	each tube
BsaI-HFv2 (R3733)^[4]	R3733L, 20 U/µL	50 units	2.5	2.75
PmeI (R0560)^[4]	R0560L, 10 U/µL	25 units	2.5	2.75
CutSmart Buffer and nuclease-free water		to 50 µL	0.16	0.176
Total			50	5.68 mix only

Put 5.16 µL of mix in each tube, then add 44.8 µL C part list.

Digest with BsaI and PmeI · 2 h

Step	Temperature	Time	Cycles
BsaI	37 °C	1 h	1^[5]
BsaI and PmeI	37 °C	1 h	1^[5]

Digest with BsaI and PmeI

Expected result

Every member of the C part list is released, about 296 bp, on the same two overhangs the destination now offers.
PmeI cuts the two external stuffers left behind, so neither can ligate back.

Troubleshooting

A member is missing from the library later: Its block was under-represented in the pool, or its own digest was incomplete. Check the pool before repeating the round.

Notes

One tube takes the whole part list: every member carries the same stuffers and differs only in its coding bases and its barcode.^[6]
The amounts are weighed at the pool's mean block length, 350 bp.

20Round 3: clean both digests up

Add 2 volumes of SPRI paramagnetic beads to each digest and elute in water.
Measure both concentrations; the ligation table asks for picomoles, not nanograms.

Expected result

AP-1 DESynR round 2, opened: 0.00512 pmol is 20 ng at 6338 bp.
C part list, released: 0.00512 pmol is 0.93 ng at 296 bp.

Troubleshooting

Too dilute for the ligation: Elute in a smaller volume, or scale the ligation up.

Notes

2 volumes here and 1 after the ligation. The two ratios differ; they are not one number used twice.

21Round 3: ligate the C part list into the library

T7 DNA Ligase: never PEG Never add PEG to a T7 ligase reaction. Blunt-end ligation is forced above 20% (w/v) PEG 6000, and StickTogether buffer already carries 7.5%; the refusal that removes escapees holds only while nothing adds more.^[1]

Pipette the two DNAs into the tube first, then the ligase, its buffer and water.
Hold at room temperature for 30 minutes.

Ligation

Reactions mix includes 10% extra

Component	Final	1 rxn (µL)	Mix for 1 (µL)
AP-1 DESynR round 2, opened^[3]	0.00512 pmol (20 ng)	1	each tube
C part list, released^[3]	0.00512 pmol (0.93 ng)	1	each tube
T7 DNA Ligase in StickTogether DNA Ligase Buffer, and nuclease-free water	to 200 µL	198	218
Total		200	218 mix only

Put 198 µL of mix in each tube, then add 1 µL AP-1 DESynR round 2, opened, 1 µL C part list, released.

Ligation

AP-1 DESynR round 2

Explore the mapClose the map

Open it in a new tab

AP-1 DESynR round 3

Explore the mapClose the map

Open it in a new tab

Round 3 of 3 joins C_JUN at position C, entering on GCAT and leaving TTCC: what it opens above, what it makes below^[7]

Expected result

Nothing visible. 0.00512 pmol of the released part list against 0.00512 pmol of the opened library is the 1:1 molar ratio.

Troubleshooting

Few colonies after this round: Check that both digests went to completion; an uncut end cannot ligate.

Notes

The 1:1 ratio is the method's, and it is a molar ratio: the shorter fragment weighs less at the same one.^[5]
The method names T7 DNA Ligase and StickTogether DNA Ligase Buffer and publishes neither the units nor the volume, so the last line of the table is the supplier's own.

22Round 3: clean the ligation up

Add 1 volume of SPRI paramagnetic beads and elute in water.
Measure the concentration.

Expected result

Enough for one electroporation: 100 ng is 0.0252 pmol at 6435 bp.

Troubleshooting

The eluate is still salty: Repeat the clean-up; salt carried into the cuvette arcs the pulse.

Notes

One volume here, not the 2 the digests took. Eluting in water rather than a buffer is what keeps the salt out of the cuvette.

23Round 3: electroporate into Endura ElectroCompetent Cells

Caution: Keep the cells and the cuvette on ice; a warm cuvette arcs.

Thaw one aliquot of Endura ElectroCompetent Cells on ice, 25 µL a pulse.
Add at most 100 ng of the purified ligation and mix without making bubbles.
Pulse at 1800 V, 600 Ω and 10 µF in a 1 mm cuvette.

Expected result

A pulse with no arc, and a time constant of 3.5 to 4.5 ms.

Troubleshooting

The cuvette arcs: The DNA carries salt. Clean it up again, elute in water, and use a fresh aliquot of cells.

Notes

The method names the instrument and none of the settings, so the program is the cells' own: it is keyed by 60242-2 and changes when the cells do. The materials table says where it was read.

24Round 3: recover and grow at 30 °C

Add recovery medium straight away and shake for the first hour.
Plate a measured dilution of the recovery on 50 µg/mL kanamycin as round 3 titre, and grow the rest in selective broth.
Plate the no-donor ligation from the same digest as round 3 no-donor control, at the same dilution.

Recovery and outgrowth at 30 °C · 13 h

Step	Temperature	Time	Cycles
Recovery, shaking	30 °C	1 h	1^[5]
Outgrowth	30 °C	12 h	1^[5]

Recovery and outgrowth at 30 °C

Expected result

At least 195,386 net colonies, scaled up from the dilution: the floor for the 0.99 chance this design asked for that none of its 13,824 distinct products is missing, equally represented.
At that count the chance a named product is missing is 7.27 × 10⁻⁷.
The round 3 no-donor control should be near empty beside it; its colonies come off the count.

Counted on round 3 titre less round 3 no-donor control Dilution factor

— net colonies.

Troubleshooting

Fewer net colonies than the count above: The round has lost library members and no later round can put them back. Electroporate more of the ligation, or run the round again.
The no-donor control is not empty: The destination was not opened to completion, or the blunt chopper missed. What grows is the round before's library, one position short, and it reaches the linkage read as truncated members.

Notes

Both steps run at 30 °C and not at 37 °C. That is a library precaution rather than an oversight.
The program carries the shorter outgrowth; 12 to 16 hours is the range the method gives.
The 0.99 was chosen for this run, and follows from the representation the screen downstream asks for. No source sets it, and nothing here defaults it; it works out at 14x this round's products.
The colony floor is what the chance of a missing member comes to over a library whose members are equally represented.^[8]
The control measures the chain the design rests on — two cuts, a blunt chopper, a ligase that refuses blunt ends and the 2x clean-up — rather than assuming it.

25Round 3: prep the library

Harvest the whole culture rather than a single colony.
Prep the plasmid and measure the concentration.

Expected result

One pool of plasmid, 6435 bp per molecule. This is the library the rounds built.
BbsI cuts it in 2 places, around the stuffer the part carried in. BsaI cuts it in 2 places, as it cut the destination: the part brought none in.

Troubleshooting

A low yield: The culture was grown too briefly, or the antibiotic was wrong for this backbone. Check the vector's own marker.

Notes

Harvesting the whole culture is what keeps the library a library; picking colonies here throws away everything not picked.

Read the library back

26Read linkage

Amplify the whole cargo out of the library the rounds built, from the vector before the first AGGA to the vector past the final TTCC, so one read carries a member's parts and its barcode block together, on Linkage read forward and Linkage read reverse from library-read-primers.tsv, which give a 1197 bp amplicon.
Sequence the amplicon as a long read.
Decode each read's block against barcodes.tsv, then read the coding bases beside it against the member that block names.

Expected result

A table from barcode combination to cargo: each read decodes to one member of each part list, and the coding bases it carries are that member's.
The 41 bp block reads C, DBD and N, each barcode separated from the last by the cloning scar TTCC, at 1402 .. 1442 of the representative construct.

Troubleshooting

A read carries fewer barcodes than there are rounds: A round did not go in. Check that round's prep against its own record before blaming the sequencing.
A combination decodes to a member the coding bases are not: That round joined a block to the wrong barcode, so barcodes.tsv no longer names what the library holds. Rebuild that round rather than carrying the table forward; no later round repairs it.

Notes

The block reads in the reverse of the order the rounds ran: each round inserted its barcode ahead of the ones already there.
None of the single-base deletions a barcode can carry leave a read another barcode of the same part list could leave, which no read can be assigned through.
This build passes the linkage read at 90.0% of reads carrying a barcode that still names its part. That is this run's own mark, not a published one.

27Read representation

Amplify across the 41 bp barcode block alone, forward from the 30 bp internal stuffer every member keeps and back from the vector past the final TTCC, on Representation read forward and Representation read reverse from library-read-primers.tsv, which give a 134 bp amplicon.
Sequence the amplicon as a short read, decode each read against barcodes.tsv, and count the reads each barcode combination gets.

Expected result

A count for each of up to 13,824 distinct combinations: how many of them are seen at all, and how evenly they are read.
A combination with no reads is a member the library has lost.
At least 99.5% of the combinations seen; a 90th/10th percentile skew ratio below 10, judged at 100 or more reads a member, which is 1,382,400 reads over 13,824 combinations.^[9]

Troubleshooting

A combination is missing: It was lost at a round or at a bottleneck since, and no later step puts it back. Read the rounds' own titre plates before re-reading this one.
The counts are heavily skewed: Members differ in length and a bottleneck can favour the short ones. The skew is priced in screen coverage: a 90th/10th ratio of 2.5 wants 200-fold, 5 wants 300-fold and 10 wants 400-fold, so a skewed library costs cells downstream rather than failing here.^[10]

Notes

Read representation again after every later bottleneck — the final assembly, and anything downstream that resamples the library. Linkage is read once; this one is read at each.
The short amplicon is what makes repeating it cheap: linkage spans the whole cargo, this spans the block.
Where you amplify the block to read it, carry any sample index on a primer rather than ligating it on: a ligated barcode spreads read counts about twofold where the same one amplified on did not.^[11]
Keep the barcodes away from where a primer anneals: a barcode 3 bp from the insert gave up to 100-fold differences in read counts, where one 34 bp away did not.^[12]

Holes

1 number this protocol would otherwise have to invent. None is filled with a guess, and each says below what it waits on: a bench, a shelf, a price, or a number nobody has published.

H23 no sourced number — the split digest, the procedure: nobody has published two enzymes run in one tube, the second added to the first's reaction with no heat kill before the beads; the one source is the paper itself. Waiting on a number nobody has published. Filled by nothing; it is the method's own.

Sources

New England Biolabs #M0318 T7 DNA Ligase product page · r.jina.ai · 2026-10-06 https://www.neb.com/en/products/m0318-t7-dna-ligase cited at reaction conditions · note 4
Endura Competent Cells manual, Lucigen / Biosearch Technologies · MA133 26Feb2018 · plain curl · 2026-10-06 cited at p. 4-5
New England Biolabs NEBioCalculator · script main-6f3020c533.js, capture 2026-08-28 · Wayback Machine · docs/research/primer-design-and-pcr.md cited at dsDNA molecular weight
New England Biolabs product specification sheets · #R3733S/L, #R3539S/L, #R0629S/L, #R0560S/L · plain curl · 2026-10-06 · docs/research/bench-numbers.md cited at #R3539L, concentration · #R0629L, concentration · #R3733L, concentration · #R0560L, concentration
Takacsi-Nagy, O. et al. (2026) Synthetic transcription factors designed by domain recombination enhance CAR T cell antitumor function. Cell 189, 1-20 · STAR Methods, METHOD DETAILS pp. e4-e5 · docs/research/protein-library-assembly.md https://doi.org/10.1016/j.cell.2026.07.054 cited at p. e4, digest · p. e4, one-pot assembly · p. e5, recovery and outgrowth
docs/projects/igga.md, the iGGA method page. The 34-base internal stuffer is the method's own, and a round excises the 30-base core of it: the two figures are the same stuffer, counted whole and counted after the entry overhang it opens on. The external stuffers lay BsaI and PmeI at the offsets their own cut offsets ask for, so every overhang is read off the DNA rather than stated. https://liuhlab.github.io/liulab-mbio/projects/igga/
liulab-mbio, Figure sources for four designs · docs/research/figure-sources.md cited at section 4, iGGA
Clarke, L. and Carbon, J. (1976) A colony bank containing synthetic ColE1 hybrid plasmids representative of the entire E. coli genome. Cell 9, 91-99 · docs/research/protein-library-assembly.md https://doi.org/10.1016/0092-8674(76)90055-6 cited at the clones a library takes
Joung, J. et al. (2017) Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening. Nat. Protoc. 12, 828-863 · docs/research/vector-qc-panel.md https://doi.org/10.1038/nprot.2017.016 cited at library representation
Imkeller, K. et al. (2020) gscreend: modelling asymmetric count ratios in CRISPR screens to decrease experiment size and improve phenotype detection. Genome Biol. 21, 53 · docs/research/vector-qc-panel.md https://doi.org/10.1186/s13059-020-1939-1 cited at Table 2
Alon, S. et al. (2011) Barcoding bias in high-throughput multiplex sequencing of miRNA. Genome Res. 21, 1506-1511 · docs/research/barcode-design.md https://doi.org/10.1101/gr.121715.111 cited at ligated against amplified barcodes
Van Nieuwerburgh, F. et al. (2011) Quantitative bias in Illumina TruSeq and a novel post amplification barcoding strategy for multiplexed DNA and small RNA deep sequencing. PLoS ONE 6, e26969 · docs/research/barcode-design.md https://doi.org/10.1371/journal.pone.0026969 cited at barcode position