Skip to content

Methods and checksMethods

How every result is reached and what it was checked against, app by app. The same as each app's Methods dialog.

Methods and checks pages

Drybench Plasmid: methods and checks

How Drybench Plasmid reaches its results: every calculator, analysis or view with its methods, the references behind them, how far each has been checked, and the app's limits.

Open Drybench Plasmid · What the levels mean · Every app's methods and checks

Map

The plasmid drawn to scale with its features, the enzyme sites in the table with dam and dcm methylation as set, and the shipped primers' sites.

Its result goes through 4 methods. The lowest level among them: Tested, not reference-checked.

Below the top level: Sequencing-primer sites (Tested, not reference-checked).

Limits that apply here

  • Enzyme sites and cut positions are from a hand-curated table of 223 enzymes (version 1, built 2026-10-06); it may differ from a vendor's list, so check NEBcutter before a critical digest. Star activity, Type IIB and nicking enzymes are not modelled.
  • Methylation blocking is called only where McClelland et al. (1994) record it; elsewhere a site reads 'may be blocked: sensitivity not recorded', which means unknown, never 'cuts'. 25 of the 82 enzymes whose site can carry a dam or dcm methylated base have no row in this app's copy of Table I, and whether Table I lists them has not been checked.
  • Only dam (GATC) and dcm (CCWGG) methylation are modelled; EcoKI, CpG and other host methylation are not, and neither is hemimethylated DNA for DpnI.
  • Feature badges are alignment matches to public records: evidence of sequence, not of function.
  • No call does not mean the feature is absent: a feature is called only if it is one of the library's 90 features and its alignment passes the thresholds; features under 30 bases are called only when found whole and exact.
  • Primer sites are exact and one-mismatch sites of the shipped primers; whether a primer primes at a mismatched site is not assessed, and a Tm is given only against a perfect match.

Methods

Restriction sites and cut positions

  • Level: Checked against a reference
  • Reference: Recognition sequences and cut positions as published facts, typed by hand (pipeline/plasmid-map/enzymes.csv); every row checked against Biopython 1.88 Bio.Restriction, whose data are REBASE release 404 (Roberts RJ et al. REBASE: a database for DNA restriction and modification. Nucleic Acids Res 2023;51:D629-D630).
  • How it was checked: Every row checked against Biopython 1.88 Bio.Restriction (enzyme data release 404) in its own notation, overhang and cut positions: 223 enzymes on pUC19 and pBR322, 446 cases, cut positions exact, with sites across the origin and 6 double digests (2026-10-06).

dam and dcm methylation of sites

  • Level: Checked against a reference
  • Reference: McClelland M, Nelson M, Raschke E. Effect of site-specific modification on restriction endonucleases and DNA modification methyltransferases. Nucleic Acids Res 1994;22:3640-3659, Table I. doi:10.1093/nar/22.17.3640 PMID 7937074
  • How it was checked: Checked against 19 cases worked by hand from McClelland et al. 1994, Table I (one per outcome, flank and origin overlaps, fully methylated DpnI and host settings, plus hemi-site sensitivity entries for other enzymes; not a hemimethylated-DpnI model), and the cited blocked examples on pUC19 (MboI) and pBR322 (EcoRII), outcomes exact (2026-10-06). Which enzymes need a Table I row (82 whose site can carry a dam or dcm methylated base, 57 of them with a row) checked by a second implementation sharing no code with the engine (2026-10-17); whether Table I lists the other 25 has not been checked.

Feature annotation from the library

  • Level: Checked against a reference
  • Reference: Local alignment with affine gaps (Smith TF, Waterman MS. J Mol Biol 1981;147:195-197; Gotoh O. J Mol Biol 1982;162:705-708), seeded by shared 12-base words and banded; match +2, mismatch -3, gap open -5, gap extend -2, this app's parameters, checked against the Python reference; a gap of k bases scores -5 - 2(k - 1), so one long gap costs less than several short ones. Identity is identical columns over alignment columns, gaps included; coverage is the feature's aligned bases over its length. The exported local_alignment_score is this score, for re-deriving a call, not a quality measure; for a feature under 30 bases no alignment runs and the score is set to twice its length, the score a whole exact alignment would have. doi:10.1016/0022-2836(81)90087-5
  • How it was checked: Checked against Biopython 1.88 PairwiseAligner with the same scoring on 26 alignments of library parts to other NCBI records: identity and coverage within 0.1 percentage points, place, strand, call and score exact; 12 synthetic inserts on each side of every threshold; no match and 0 partial calls on 10 random sequences of 10,000 bp and both golden plasmids reversed (2026-10-06). Library 2 (2026-10-17): all 4,005 pairs of its parts aligned against each other with the same scoring, 66 of them reaching the thresholds and each published with how it is shown; no call on 12 random controls; what is shown at a locus, the parts named beside it and the inside labels exact on 14 sequences.

Sequencing-primer sites

  • Level: Tested, not reference-checked
  • Reference: Exact and one-mismatch matches of the shipped primers on both strands; a mismatch is reported with its distance from the primer's 3′ end, and nothing is claimed about whether such a site primes.
  • How it was checked: Each of the 6 shipped primers found exactly at its NCBI binding site at build and every week; one-mismatch places unit-tested (2026-10-06).

Features

Library features found on both strands and across the origin, with the alignment's numbers, and the features you added or edited.

Its result goes through 3 methods. The lowest level among them: Tested, not reference-checked.

Below the top level: Sequencing-primer sites (Tested, not reference-checked).

Limits that apply here

  • Feature badges are alignment matches to public records: evidence of sequence, not of function.
  • No call does not mean the feature is absent: a feature is called only if it is one of the library's 90 features and its alignment passes the thresholds; features under 30 bases are called only when found whole and exact.
  • Primer sites are exact and one-mismatch sites of the shipped primers; whether a primer primes at a mismatched site is not assessed, and a Tm is given only against a perfect match.

Methods

Feature annotation from the library

  • Level: Checked against a reference
  • Reference: Local alignment with affine gaps (Smith TF, Waterman MS. J Mol Biol 1981;147:195-197; Gotoh O. J Mol Biol 1982;162:705-708), seeded by shared 12-base words and banded; match +2, mismatch -3, gap open -5, gap extend -2, this app's parameters, checked against the Python reference; a gap of k bases scores -5 - 2(k - 1), so one long gap costs less than several short ones. Identity is identical columns over alignment columns, gaps included; coverage is the feature's aligned bases over its length. The exported local_alignment_score is this score, for re-deriving a call, not a quality measure; for a feature under 30 bases no alignment runs and the score is set to twice its length, the score a whole exact alignment would have. doi:10.1016/0022-2836(81)90087-5
  • How it was checked: Checked against Biopython 1.88 PairwiseAligner with the same scoring on 26 alignments of library parts to other NCBI records: identity and coverage within 0.1 percentage points, place, strand, call and score exact; 12 synthetic inserts on each side of every threshold; no match and 0 partial calls on 10 random sequences of 10,000 bp and both golden plasmids reversed (2026-10-06). Library 2 (2026-10-17): all 4,005 pairs of its parts aligned against each other with the same scoring, 66 of them reaching the thresholds and each published with how it is shown; no call on 12 random controls; what is shown at a locus, the parts named beside it and the inside labels exact on 14 sequences.

Sequencing-primer sites

  • Level: Tested, not reference-checked
  • Reference: Exact and one-mismatch matches of the shipped primers on both strands; a mismatch is reported with its distance from the primer's 3′ end, and nothing is claimed about whether such a site primes.
  • How it was checked: Each of the 6 shipped primers found exactly at its NCBI binding site at build and every week; one-mismatch places unit-tested (2026-10-06).

Oligo melting temperature by nearest neighbour

  • Level: Checked against a reference
  • Reference: SantaLucia J Jr. A unified view of polymer, dumbbell, and oligonucleotide DNA nearest-neighbor thermodynamics. Proc Natl Acad Sci USA 1998;95:1460-1465 (Table 2; read 6 October 2026); concentration models after SantaLucia J Jr, Hicks D. Annu Rev Biophys Biomol Struct 2004;33:415-440. doi:10.1073/pnas.95.4.1460
  • How it was checked: The sequence module's Tm, checked against Biopython 1.88 within 0.1 °C on 45 oligos (the sequence app's fixtures); here against a perfect match only.

Digest

Choose enzymes and see the fragments and a schematic gel, or let the picker find a digest that tells the correct construct from an empty or reversed one.

Its result goes through 5 methods. The lowest level among them: Tested, not reference-checked.

Below the top level: Schematic gel (Tested, not reference-checked).

Limits that apply here

  • Enzyme sites and cut positions are from a hand-curated table of 223 enzymes (version 1, built 2026-10-06); it may differ from a vendor's list, so check NEBcutter before a critical digest. Star activity, Type IIB and nicking enzymes are not modelled.
  • Methylation blocking is called only where McClelland et al. (1994) record it; elsewhere a site reads 'may be blocked: sensitivity not recorded', which means unknown, never 'cuts'. 25 of the 82 enzymes whose site can carry a dam or dcm methylated base have no row in this app's copy of Table I, and whether Table I lists them has not been checked.
  • Only dam (GATC) and dcm (CCWGG) methylation are modelled; EcoKI, CpG and other host methylation are not, and neither is hemimethylated DNA for DpnI.
  • The gel is schematic: band order and spacing by the logarithm of size only, no millimetre or Rf positions; fragments under 100 bp are not drawn; a band holds the fragments within 5 percent of its largest (a judgement), which may not resolve; the forms of uncut DNA are not modelled.

Methods

Restriction sites and cut positions

  • Level: Checked against a reference
  • Reference: Recognition sequences and cut positions as published facts, typed by hand (pipeline/plasmid-map/enzymes.csv); every row checked against Biopython 1.88 Bio.Restriction, whose data are REBASE release 404 (Roberts RJ et al. REBASE: a database for DNA restriction and modification. Nucleic Acids Res 2023;51:D629-D630).
  • How it was checked: Every row checked against Biopython 1.88 Bio.Restriction (enzyme data release 404) in its own notation, overhang and cut positions: 223 enzymes on pUC19 and pBR322, 446 cases, cut positions exact, with sites across the origin and 6 double digests (2026-10-06).

Digest fragments

  • Level: Checked against a reference
  • Reference: Fragments between consecutive top-strand cuts, across the origin of a circular plasmid; lengths in base pairs on the top strand, as Biopython and NEBcutter report them.
  • How it was checked: Fragments checked against Biopython 1.88 Bio.Restriction on 446 single digests and 6 double digests of pUC19 and pBR322, lengths exact (2026-10-06).

dam and dcm methylation of sites

  • Level: Checked against a reference
  • Reference: McClelland M, Nelson M, Raschke E. Effect of site-specific modification on restriction endonucleases and DNA modification methyltransferases. Nucleic Acids Res 1994;22:3640-3659, Table I. doi:10.1093/nar/22.17.3640 PMID 7937074
  • How it was checked: Checked against 19 cases worked by hand from McClelland et al. 1994, Table I (one per outcome, flank and origin overlaps, fully methylated DpnI and host settings, plus hemi-site sensitivity entries for other enzymes; not a hemimethylated-DpnI model), and the cited blocked examples on pUC19 (MboI) and pBR322 (EcoRII), outcomes exact (2026-10-06). Which enzymes need a Table I row (82 whose site can carry a dam or dcm methylated base, 57 of them with a row) checked by a second implementation sharing no code with the engine (2026-10-17); whether Table I lists the other 25 has not been checked.

Schematic gel

  • Level: Tested, not reference-checked
  • Reference: Band order and spacing by the logarithm of size only; no millimetre or Rf positions, since no calibrated mobility model is used. Fragments under 100 bp are not drawn. Sizes taken largest first, a fragment joins a band only if it is within 5 percent (inclusive, a judgement) of that band's largest fragment, so a band's fragments are within 5 percent of each other; they may not resolve and are drawn as one. Two lanes are compared band by band through each band's largest fragment. Supercoiled, nicked and linear forms of uncut DNA are not modelled.
  • How it was checked: Grouping checked against 11 hand-worked cases with exact fractions (gel rules 2), order and log spacing unit-tested (2026-10-17); no oracle exists for band positions, so none is drawn.

Diagnostic digest picker

  • Level: Checked against a reference
  • Reference: Single and double digests compared on the plasmid as given, with the insert removed (empty) and with the insert reversed; a digest is offered when the schematic gel tells all three apart and cuts each of them; the verdicts are read from bands grouped under gel rule 2.0.0; the pool and the ranking are this app's judgement.
  • How it was checked: Checked against the oracle's verdicts on the correct, empty and reversed trio (EGFP in pUC19's SmaI site both ways, and pUC19 empty) under gel rule 2.0.0 (trio v2, 2026-10-17): the pool of 95 enzymes and all 4,157 digests judged agree, bands included, 1,710 of which tell all three apart (unmethylated DNA). 43 of 4,157 digests of the EGFP-in-pUC19 trio change their bands under gel rule 2.0.0; no picker verdict on that trio changes, and the digests telling all three apart are the same set as under gel rule 1.0.0 (trio v1, 2026-10-06).

Sequence

The sequence, 60 bases to a line, with the features under it.

No method record is kept for this view, so it carries no level.

Numbers

From this plasmid's own sequence: its double-stranded weight, and the amounts that follow from it.

Its result goes through 2 methods. The lowest level among them: Checked against a reference.

Methods

ng, pmol, copies and ligation amounts

  • Level: Checked against a reference
  • Reference: pmol = ng × 1000 / MW; copies = ng × 10⁻⁹ × N_A / MW, with N_A = 6.02214076 × 10²³ per mole (exact, SI 2019); insert ng = vector ng × insert bp / vector bp × ratio. MW of the double-stranded plasmid from its own sequence (both strands, no end groups when circular) with the CIAAW 2021 atomic weights.
  • How it was checked: Checked against exact rational arithmetic on 11 cases (pmol, copies per µL, ng for copies, ligation, the 650 Da per bp rule and the weights of pUC19 and pBR322), to full double precision (2026-10-06).

Nucleic acid molecular weight from the sequence

  • Level: Checked against a reference
  • Reference: Residue formulas from the nucleoside 5′-monophosphates (PubChem, read 6 October 2026) less water, summed with the IUPAC CIAAW 2021 atomic weights; 5′-OH and 3′-OH ends as a neutral free acid.
  • How it was checked: Checked against exact rational arithmetic on 11 cases (pmol, copies per µL, ng for copies, ligation, the 650 Da per bp rule and the weights of pUC19 and pBR322), to full double precision (2026-10-06).

Explainers

Short, cited explanations of what is on a plasmid map, each linked to the stop of the cell tour it illustrates.

No method record is kept for this view, so it carries no level.

Assembly

Restriction cloning, Golden Gate and Gibson-style designs from this plasmid and an insert, with the ends checked and the product's sequence written out; a design, not evidence of a reaction.

Its result goes through 6 methods. The lowest level among them: Tested, not reference-checked.

Below the top level: Fixed terminal primers and one-sided Gibson overlap tails (Tested, not reference-checked).

Limits that apply here

  • Assembly is sequence design, not evidence of a reaction. Restriction and Golden Gate refuse overlapping staggered cleavage intervals; restriction selects the longest enzyme-ended fragments and reports compatible orientations. Golden Gate supports BsaI and BsmBI. Gibson takes 2–6 ordered linear fragments with exact terminal homology of at least 15 bp and no repeated homology on either strand. Products are limited to 50,000 bp. No repair, efficiency or experimental-success prediction is made.

Methods

Double-stranded fragment ends and compatibility

  • Level: Checked against a reference
  • Reference: A restriction cut is a pair of positions, the top-strand bases before the top-strand and bottom-strand cuts; the exposed overhang is the bases between them, a 5′ overhang when the top cut is 5′ of the bottom and a 3′ overhang when 3′ of it, blunt when they coincide. Two ends anneal only when both are 5′ overhangs, both 3′ or both blunt and one overhang read 5′→3′ is the reverse complement of the other, decided on the exposed bases and polarity, never on the enzyme name or a top-strand coordinate. A fragment and its reverse complement are one molecule: written out from the other strand, its two ends swap sides and keep their own overhang bases and polarity, only the protruding strand changing, so it anneals in the same places whichever strand it was supplied on. Paired cores exclude both staggered overhangs; coincident physical cuts are deduplicated and overlapping staggered cleavage intervals are refused. Circular strand cuts retain their signed separation before wrapping, so an origin change cannot change end polarity or overhang length. No end is repaired, filled in, phosphorylated, blunted or mutated.
  • How it was checked: Checked against 5 cut cases and 8 compatibility cases worked by hand (EcoRI 5′ AATT, KpnI 3′ GTAC, a blunt cut and two non-palindromic overhangs, with BamHI/BglII compatibility and polarity and length mismatches), and 3 flip cases (a Type IIS part with distinct 5′ overhangs, non-palindromic 3′ overhangs and a blunt fragment), each written out base by base and re-derived from a two-strand drawing: every overhang, polarity and compatibility verdict exact, and writing a fragment out from the other strand keeps each end's own bases so one molecule anneals in the same places either way (10 October 2026).

Restriction cloning

  • Level: Checked against a reference
  • Reference: A vector and an insert are each cut by the chosen enzymes (the restriction and digest methods above, under the host's dam and dcm marks); the chosen backbone and insert fragments are ligated into a circle when their exposed ends anneal, decided on the exposed bases and polarity alone (the double-stranded end method above), never on the enzyme name or a coordinate. The insert is tried both ways and every orientation that anneals is shown, so a symmetric or blunt join is surfaced as ambiguous, not resolved silently. Material is conserved: the product length is the sum of the fragments' top-strand lengths less each shared overhang counted once. The products shown are the one-insert circles whose ends anneal: vector self-ligation, circles carrying more than one insert and concatemers are not enumerated, so the list is not every product a ligation can give. The worked example takes the longest enzyme-ended fragment on each side by default. No end is repaired, filled in, phosphorylated, blunted or mutated; no ligation efficiency, vector:insert ratio or transformation outcome is given; a cut off a linear end refuses. The product carries a segment map: which input each run of product bases came from and where, a shared overhang belonging to both the inputs that hold it. A designed product is not evidence that the reaction occurred.
  • How it was checked: Checked against 3 clones and 2 refusals worked by hand: a directional EcoRI/HindIII clone, single-enzyme (BamHI) and blunt (SmaI) clones that permit both insert orientations and so are shown as two products, incompatible ends refused and a Type IIS cut off a linear end refused. Each hand product is round-tripped by Biopython 1.88 Bio.Restriction (it digests back into the insert and backbone cores); material conserved on every product (10 October 2026).

Golden Gate assembly (BsaI, BsmBI)

  • Level: Checked against a reference
  • Reference: A Type IIS enzyme (BsaI GGTCTC(1/5) or BsmBI CGTCTC(1/5), offsets checked against Biopython Bio.Restriction) cuts outside its recognition sequence, so a part flanked by two inward-pointing sites is released with two defined 5′ overhangs while the recognition sites stay on the discarded flanks; the opened destination backbone and the parts assemble by their exposed overhangs into one scarless circle, decided on the overhangs and polarity alone. Each part is matched in either orientation, since a part and its reverse complement are one double-stranded molecule, so the assembly does not depend on which strand a part was supplied on; the destination backbone is the reference frame. Overhangs must be 5′, non-palindromic and all distinct, and must close into a single cycle; a palindromic overhang, a part that could follow in either orientation, a duplicate or reverse-complement overhang conflict, an unresolved internal site or a cut off a linear end is refused, never a warning to click past. Recognition sites are scanned on physical strand sequences, including the final circular product and every assembled junction; a newly created site is refused. No part is domesticated and no coding sequence changed automatically; no fidelity or ligation-fidelity figure is given. Material is conserved. The product carries a segment map: which input each run of product bases came from and where, a shared overhang belonging to both the inputs that hold it. A designed product is not evidence that the reaction occurred.
  • How it was checked: Checked against 3 assemblies and 4 refusals: a two-part BsaI and a one-part BsmBI assembly whose parts are built around a chosen scarless product, and the same two-part design with one part supplied on the other strand giving the same product, with Bio.Restriction 1.88 cut positions showing the released overhangs, and no recognition site surviving in the product; a palindromic overhang, a duplicate overhang, an unresolved internal site and a Type IIS cut off a linear end each refused (10 October 2026).

Gibson-style assembly

  • Level: Checked against a reference
  • Reference: A user-ordered plan of 2 to 6 fragments joined on exact terminal overlaps: the overlap between two consecutive fragments is the longest suffix of one that is a prefix of the next, found base for base, collapsed once; the plan closes into a circle or is left linear, within a 50,000 bp product ceiling. The method depends on terminal homology long enough for the recessed strands to anneal (Gibson et al. 2009, Nat Methods 6:343–345, doi 10.1038/nmeth.1318; the commonly used working range is roughly 15 to 40 bases), so an overlap shorter than 15 bp withholds the product: the floor is a limit of applicability, not a recommended length. Sequence planning on double-stranded homology, not a thermodynamic model: no overlap length is recommended or called optimal and no melting temperature is used; fragments are taken as supplied and nothing is flipped. A missing or contradictory overlap, or an overlap that occurs more than at its two designed termini when the overlap and its reverse complement are counted across all fragments (repeated homology on either strand, which would permit an alternative or inverted assembly), withholds the product rather than resolving it silently. Every fragment must retain a non-overlapping core after both terminal overlaps; fully consumed fragments, bridge oligos and overlapping junction assignments are outside this planner and withhold the product. Material is conserved: the product length is the sum of the fragment lengths less each collapsed overlap once. The product carries a segment map: which input each run of product bases came from and where, a collapsed overlap belonging to both the fragments that hold it. A designed product is not evidence that the reaction occurred.
  • How it was checked: Checked against 2 assemblies and 5 refusals: a three-fragment circular plan with 20 bp overlaps and a two-fragment linear plan with an overlap exactly at the 15 bp floor, whose fragments are cut from a chosen product and extended by exact overlaps, with the product re-derived by an independent concatenation, material accounted base by base, and reverse-complement and rotation giving the same product; a missing overlap, an 8 bp and a 14 bp overlap below the floor, and repeated homology on the supplied strand and on the other strand only each withheld (10 October 2026).

Fixed terminal primers and one-sided Gibson overlap tails

  • Level: Tested, not reference-checked
  • Reference: For each explicitly oriented linear template, use its first N bases as the forward annealing region and the reverse complement of its last N bases as the reverse annealing region. The forward tail is empty. The reverse 5′ tail is the reverse complement of the next oriented template’s first H bases; the last linear fragment has no tail. Require one exact full-template PCR product at the intended sites, then run the existing exact-overlap Gibson checker. Every observed junction must equal the requested H-base prefix: otherwise withhold the product. The final sequence is the oriented templates concatenated, with no raw template bases removed. N is 8–40 nt, H is 15–40 bp; these are supported bounds, not recommendations. Two to six fragments and a 50,000 bp final product are supported. Tm is retained for each annealing region only, with its original model, conditions, applicability notes or refusal. No optimisation, amplification, specificity or reaction-success claim is made.
  • How it was checked: Checked against independent Python string-algebra fixtures gibson-primers-v1/v2 in the repository, including literal oligos and the 50,000 bp boundary. This is sequence-design checking, not experimental reaction validation.

Oligo melting temperature by nearest neighbour

  • Level: Checked against a reference
  • Reference: SantaLucia J Jr. A unified view of polymer, dumbbell, and oligonucleotide DNA nearest-neighbor thermodynamics. Proc Natl Acad Sci USA 1998;95:1460-1465 (Table 2; read 6 October 2026); concentration models after SantaLucia J Jr, Hicks D. Annu Rev Biophys Biomol Struct 2004;33:415-440. doi:10.1073/pnas.95.4.1460
  • How it was checked: The sequence module's Tm, checked against Biopython 1.88 within 0.1 °C on 45 oligos (the sequence app's fixtures); here against a perfect match only.

SpCas9 guide sites

Find every SpCas9 guide (20 base protospacer, NGG PAM) on both strands, with GC, homopolymer run and the exact within-sequence count; no score or genome claim.

It has a page of its own: SpCas9 guide sites.

Its result goes through 1 method. The lowest level among them: Checked against a reference.

Methods

SpCas9 guide sites

  • Level: Checked against a reference
  • Reference: Every 20 nt protospacer immediately 5′ of an NGG PAM (SpCas9), on both strands, matched base for base; a PAM across the origin of a circular sequence is found once. Each candidate carries its coordinates and strand, the protospacer, the PAM, the GC fraction (G+C over the 20 protospacer bases), the longest homopolymer run and the exact count of other occurrences of the protospacer within the pasted sequence on both strands. No on-target or off-target score of any kind, no mismatch-tolerant search and no genome-wide or specificity statement: the count is of exact matches inside the sequence you pasted, it is not an off-target assessment and says nothing about the genome. SpCas9 (NGG, 20 nt) only. Candidates are ordered by position.
  • How it was checked: Checked against an independent enumeration on 5 sequences, 21 SpCas9 guides in all: one regular-expression search for a protospacer and its PAM, run over the sequence for the + strand and over Biopython 1.88's reverse complement for the − strand, each hit's position mapped back to top-strand coordinates, with the sequence doubled first when circular. Every position, strand, protospacer, PAM, GC, homopolymer run and within-sequence occurrence count exact, with one case worked by hand base by base (the script refuses to write unless its search agrees with the hand work), and a tandem repeat, an opposite-strand other site and a PAM across a circular origin among them (10 October 2026).

Data sources

Limits

Not medical advice. Nothing here is advice about any medicine, treatment or condition, and nothing here is an assessment of any person's variant, sample or molecule. Decisions made with it are yours.

  • For research and education only, and not medical advice: plasmid maps and sequence assembly designs. A designed product is not evidence that the reaction occurred; nothing here predicts expression or how well a reaction will work.
  • Enzyme sites and cut positions are from a hand-curated table of 223 enzymes (version 1, built 2026-10-06); it may differ from a vendor's list, so check NEBcutter before a critical digest. Star activity, Type IIB and nicking enzymes are not modelled.
  • Methylation blocking is called only where McClelland et al. (1994) record it; elsewhere a site reads 'may be blocked: sensitivity not recorded', which means unknown, never 'cuts'. 25 of the 82 enzymes whose site can carry a dam or dcm methylated base have no row in this app's copy of Table I, and whether Table I lists them has not been checked.
  • Only dam (GATC) and dcm (CCWGG) methylation are modelled; EcoKI, CpG and other host methylation are not, and neither is hemimethylated DNA for DpnI.
  • Feature badges are alignment matches to public records: evidence of sequence, not of function.
  • No call does not mean the feature is absent: a feature is called only if it is one of the library's 90 features and its alignment passes the thresholds; features under 30 bases are called only when found whole and exact.
  • The gel is schematic: band order and spacing by the logarithm of size only, no millimetre or Rf positions; fragments under 100 bp are not drawn; a band holds the fragments within 5 percent of its largest (a judgement), which may not resolve; the forms of uncut DNA are not modelled.
  • A ladder lane draws its sheet's sizes as listed; the gel gives no band positions, no sizing by interpolation, no match with a ladder band, no migration at any agarose percentage and no supercoiled or nicked behaviour.
  • Primer sites are exact and one-mismatch sites of the shipped primers; whether a primer primes at a mismatched site is not assessed, and a Tm is given only against a perfect match.
  • SnapGene .dna files are read-only, and a file with a block this reader does not know is refused; the reader is checked on synthetic files built from NCBI records.
  • Opening a public accession sends it, and your IP address, to NCBI; pasted sequences and files do not leave your browser.
  • Plasmids up to 50,000 bp. A link carries a plasmid up to about 5,500 bp; a larger plasmid travels as a GenBank file or by Send to on this browser, and your feature edits travel only in a GenBank file.
  • Not biosafety advice. Follow your institution's rules for viral vectors and selection markers.
  • No offline use: the pages are served by the site, and nothing is installed.
  • Assembly is sequence design, not evidence of a reaction. Restriction and Golden Gate refuse overlapping staggered cleavage intervals; restriction selects the longest enzyme-ended fragments and reports compatible orientations. Golden Gate supports BsaI and BsmBI. Gibson takes 2–6 ordered linear fragments with exact terminal homology of at least 15 bp and no repeated homology on either strand. Products are limited to 50,000 bp. No repair, efficiency or experimental-success prediction is made.