Drybench Calc.

What the app shows, how it works and how it was built, in the words of its own write-up, followed by the parameter defaults and the data sources with their licences.

What it does

One live page, the bench sheet, holds the calculations a bench day needs: the mass to weigh for a molarity, the volume for a mass you actually weighed, a C1V1 dilution with units that convert as you change them, a serial-dilution plan with the volumes per tube and the DMSO carried from the stock, a protein concentration from an A280 reading with the extinction coefficient computed from the sequence, a buffer's pKa at the working temperature, and a multi-component recipe table. Every input has its unit beside it and the answer updates as you type, rounded once at display; the link keeps the inputs, never the results, so a bookmark reopens the same sheet and recomputes it.

Salt forms, hydrates and purity are handled where they bite. A compound that arrives from the molecule app carries the formula of the whole input and of its parent, so an imatinib mesylate stock is weighed on 589.72 g/mol, not the parent's 493.62, and the sheet says what the shortfall would have been. A formula weight typed from the vial label or a certificate of analysis overrides every computed value, and a purity below 100 percent scales the mass. A curated table of 75 common reagents with their hydrate forms, PubChem identifiers and formula weights fills the field with one pick.

How it works

Every calculation runs in your browser from three pure modules the suite shares: units (SI prefixes as exact powers of ten, one base unit per quantity, so chained conversions never accumulate error), dilution (C1V1, stocks with purity and the parent equivalent, serial plans and recipes, and the buffer table from Goldberg, Kishore and Lennen 2002 with the Clarke and Glew temperature equation) and seq (the average molecular weight and the Pace 1995 extinction coefficient of a protein sequence, shared with the protein app so both show the same number). Formula weights are exact sums of the IUPAC CIAAW 2021 standard atomic weights, using the conventional value for the elements given as an interval, and are rounded once when shown.

The scientific methods run locally: the app contacts no host, needs no account for scientific use; optional Bench sign-in uses session cookies. The reagent table is a static file served with the app. A stock, a dilution plan, a buffer or a recipe you pin, send or add to a study is kept in this browser as a recipe object, with its inputs and the method versions it was made with, so it recomputes when reopened.

How it is checked

Each module is compared with a named external oracle on a published fixture set, never with its own output: 28 dilution and molarity cases worked exactly in rational arithmetic, including unit chains and the pipetting-flag edges; 58 formula weights of salts, hydrates and their parents with the formula from RDKit 2026.03.6 and the weight as a CIAAW 2021 sum in a separate decimal implementation, with PubChem's identifiers recorded; 15 buffers at 4, 25 and 37 °C against Goldberg's equation in NumPy; and 22 protein sequences from UniProt against Biopython's ProtParam for ε280, which agrees with the ExPASy ProtParam outputs recorded for nine of them. The fixture files, their hashes and the script that wrote them are published on the verification page.

Equations and defaults

Units

SI prefixes as exact powers of ten with one base unit per quantity (mol/L, L, g, g/mol); the micro sign and the Greek mu are read as one character; a value is converted once and rounded once when shown, half away from zero.

Molarity and stocks

n = C·V and m = n·M / purity; the volume for a weighed mass is V = m·purity / (C·M); the parent-equivalent mass of a salt or hydrate is m·MW(parent) / FW(form).

C1V1

C1·V1 = C2·V2 with the diluent V2 − V1; a stock weaker than the final concentration is refused rather than solved.

Sample normalisation for gel loading

The Dilution calculator's second view (method sample-normalisation 1.0.0). For a load volume V per well and a loading-dye stock factor k (0 for no dye, otherwise above 1), the dye is D = V/k and A = V − D is left for sample and diluent; the sample volume is s = m/c for a target mass m, or C·V/c for a target concentration C, with each sample's concentration c in mg/mL as entered. A sample with s > A is withheld with no volumes, and the largest amount it can deliver (c·A µg, or c·A/V mg/mL) is shown; otherwise the diluent is A − s and the planned mass s·c on unrounded volumes. A positive transfer below 2 µL is flagged, never refused. A concentration that is not a positive number makes its row invalid; a load or target that is not above zero, or a dye factor that is neither 0 nor above 1, refuses the batch. The view opens empty unless concentrations arrive from a standard curve in Assay; its worked example of five illustrative lysates loads only when chosen, carries a banner until cleared, and never reaches a worklist. It does not check the assay, the protein or the well's capacity, and volumes are taken as additive.

Serial dilution

Each tube holds R·W·f/(f − 1) before it passes V/f to the next, the last tube is made the same way and keeps V/f over after its wells (2.0.0; 1.0.0 made it to R·W), and the top tube takes V·C(top)/C(stock) from the stock; the solvent fraction at a point is the stock's fraction scaled by C(point)/C(stock).

Plate preparation sheet

An optional section of the serial dilution (method plate-sheet 1.0.0) that plans the plate the series fills. Each tube keeps batch(R, W, p, D) = R·W·(1 + p/100) + D for R replicate wells of W with overage p and dead volume D (overage on the bulk, per well, or both); a tube is made to keep·f/(f − 1) and passes tube/f to the next, as the constant-transfer serial dilution does; the top tube is drawn from the stock, or from an optional intermediate at or above the top point's concentration and below the stock's. The vehicle is diluent at the top point's carrier fraction, made from neat carrier for its own wells. Dead volume is charged once per prepared source, never per well. Dose positions come from the same placement Assay reads, so the two apps cannot disagree; vehicle, blank and positive-control lines are typed well ranges, and a positive control is listed with no concentration because you supply it. A plan that cannot be made gives no sheet and names why: an overage outside 0 to 1,000 percent, a negative dead volume, a series that does not plan, a stock weaker than the top point, an intermediate outside that range, a well range or start well that does not read or lies off the plate, wells that overlap or do not fit, a transfer below the minimum, a carrier above the maximum, or not enough stock. Everything on it is planned, not measured; volumes are taken as additive, and repeated wells share a stock and are not independent replicates. Print plate plan shows the same plan on paper with its map and every row, and adds no arithmetic.

Master mix

B = N × M × (1 + extra% / 100) + dead volume. Each shared component scales by B / M; separate additions scale by N. Final concentration refers to the completed reaction, including separate additions. One common liquid formulation and one mix vessel; volumes are treated as additive. The planner does not assess reagent compatibility, enzyme activity or reaction conditions. Separate samples must use the same entered volume; make another plan for a different formulation. Overage and dead volume default to zero; a minimum transfer is optional and applies to actual handling volumes. Method version 1.0.0; primary references read 9 October 2026: Thermo Fisher MAN0025706 B pp. 12–13, INTEGRA dead-volume guidance and Eppendorf small-volume pipetting guidance.

Bench flags

A pipetting step below 2 µL and a solvent fraction above 0.5 percent are flagged; both are bench conventions adopted on 6 October 2026, not instrument specifications.

Liquid-handler worklists

The plate sheet and sample normalisation write Tecan worklists: A;RackLabel;;;Position;;Volume records, then D; and W;, with positions numbered rear to front, then left to right (A1 = 1, B1 = 2, A2 = 9 on 96 wells), volumes in µL at two decimals with a warning when rounding moves one by more than 0.1 percent, and comment lines that say the file is a plan. The plate sheet's file covers only the dispense, from a tube rack, Tubes, to the Plate in the Dispense table's order; the tubes are prepared by hand. Normalisation's file draws diluent, then dye, then sample for each row, from Diluent, Dye and Samples to a 96-well Prep plate, and leaves out a zero volume. It is refused for a batch with a withheld or invalid row, for the worked example, or for more than 96 samples. A refused file is never offered for download. Neither offers an Echo pick list: both draw from tubes, and an Echo draws only from source-plate wells.

Bench checklist

Opened from C1V1, serial dilution, master mix, sample normalisation and the printable protocol. It lists the plan's own steps and volumes, with no added procedure, beside large tick boxes, and carries the plan's own flags (the 2 µL convention, the master mix's minimum transfer). Ticks are kept while it is open and cleared when it closes: a tick is a reminder on the screen, not a record that a step was done. The optional timer counts down from a stored end time.

Formula weights

Exact sums of the IUPAC CIAAW 2021 standard atomic weights held as integers in units of 10⁻⁴ g/mol, conventional values for the interval elements, hydrates as the anhydrous formula plus n × 18.015; imatinib 493.615 and its mesylate 589.715, shown as 493.62 and 589.72.

pKa at temperature

Goldberg, Kishore and Lennen 2002, equation 12: pK(T) = pK(Tr) − (ΔrH°/(R ln 10))(1/Tr − 1/T) − (ΔrCp°/(R ln 10))(Tr/T − 1 + ln(T/Tr)), with R = 8.31451 J/(K mol) and Tr = 298.15 K, from the paper's recommended values at zero ionic strength; the linear slope −ΔrH°/(R ln 10 Tr²) is shown as the approximation it is.

ε280 and A280

Pace et al. 1995: 5,500 per Trp, 1,490 per Tyr and 125 per cystine, in both cysteine conventions; c = A/(ε·l) and mg/mL = c·MW, with the average molecular weight from the ExPASy ProtParam residue masses the protein app uses.

A260

50, 40 and 33 µg/mL per absorbance unit for double-stranded DNA, RNA and single-stranded DNA, labelled as rules of thumb; no nearest-neighbour ε260 in this version.

Reagent table

75 reagents hand-typed with their formulas and hydrate waters, each row's PubChem CID, InChIKey and formula read on 6 October 2026 and checked against the formula; pKa values are the Goldberg entries each buffer names, typed once.

Data sources, licences and versions

Times on screen are local time with the UTC offset; times in exports are UTC. Dates with no time are shown at the precision the source gave them.

The Export menu offers the bench sheet as CSV (every input and answer, units in the headers, the method and constants table per row, with the provenance sidecar), the serial-dilution plan as TSV with units in the column headers (the table the assay analyser reads), the recipe as JSON (inputs, components, method versions and constants tables, never results), the bench sheet as a printed page with the method block, a stock label with a QR code of the calculation's link, the plate preparation sheet as CSV and as a printed plate plan with its map, and Tecan worklists for the plate sheet and for sample normalisation, whose plan also copies as a tab-separated or Markdown table. A worklist is a plan for the robot to follow, written only when every row is planned. Copy link puts your inputs after the # of the address; the method block copies as text.

Release gate: before the app leaves the preview, its exports and printed reports will carry the data versions, parameters, limits and a timestamp needed to reproduce a result, as the regulatory position says. This section will describe each format as it ships.

Each module below is a set of automated tests in the app's source code, with the expected results written into the tests. They check the app's logic against recorded inputs and stored reference records; none of them calls a live service.

src/lib/units/units.test.ts and formula-mass-fixtures.test.ts

Unit conversions and spellings; every formula weight in fixtures/formula-mass.json against the CIAAW sum, exactly, with RDKit's own weight recorded beside each organic.

src/lib/dilution/dilution.test.ts and fixtures.test.ts

C1V1, stocks, the parent equivalent, serial plans with their flags and recipes against fixtures/dilution.json to nine significant figures, and pK at three temperatures against fixtures/buffers.json within 0.01 and 0.0001.

src/lib/dilution/normalization.test.ts

Sample normalisation against pipeline/lab-calculators/fixtures/normalization-v1.json (12 cases, written in Python fractions by pipeline/lab-calculators/normalization/reference.py from the method, never from the engine): volumes, masses and limits at relative 1e−12; statuses, flags and refusals exactly.

src/lib/dilution/plate-sheet-oracle.test.ts

The plate preparation sheet against src/lib/dilution/fixtures/plate-sheet-2.json, written in Python fractions by pipeline/lab-calculators/plate_sheet_reference_2.py: values at relative 1e−9; wells, roles, sources and refusals exactly. The oracle places dose wells by its own reading of the placement rule and checks one 96-well and one 384-well layout against wells written out by hand; the engine's wells, from the shared placement, must match exactly.

src/lib/export/liquid-handlers/liquid-handlers.test.ts

The Tecan writer against the record layout, the position numbering and five example lines in liquid-handler-formats-v1.json, transcribed on 17 October 2026 from TeselaGen's Tecan worklist page and robotools' example worklists. The Echo writer against the column header, the nL unit and the droplet sizes recorded there from TeselaGen's Echo worklist page and two Beckman Coulter product sources. Both refusals return no file text.

src/components/tools/lab-calculators/plate-sheet-worklist.test.ts and normalization-worklist.test.ts

The plate sheet's worklist on its oracle's first planned case: every dispense row becomes one aspirate, dispense and wash triple at its numeric positions in µL. The normalisation worklist on every planned case of its oracle: each sample keeps its own source and destination position, a batch with a withheld or invalid row writes no file, and a 97th sample refuses the file.

src/lib/dilution/master-mix.test.ts and src/apps/lab-calculators/master-mix-workspace.test.ts

Independently hand-worked preparation quantities, volume/composition conservation, explicit shared and separate allowances, mixed units, impossible input refusal, exact complete/incomplete recipe resume and immutable result retention. This is arithmetic verification; it does not validate a laboratory protocol or pipette.

src/lib/seq/protein-mass.test.ts and extinction-fixtures.test.ts

ε280 and the molecular weight against the recorded ExPASy ProtParam outputs and Biopython's ProtParam on fixtures/extinction.json, exactly; a guard that the constants live in src/lib alone.

src/lib/fixture-hashes.test.ts and pipeline/lab-calculators/verify-reagents.test.ts

Every fixture file matches the SHA-256 the oracle script recorded and its published copy; every reagent's weight recomputes from its formula and agrees with PubChem's at the precision PubChem prints.

pipeline/lab-calculators/reference.py

The oracle script (Python 3.11 with the pinned requirements.txt): rational and decimal arithmetic, RDKit 2026.03.6, NumPy 2.4.6 and Biopython 1.85; it writes the fixtures and their hashes and copies them under public/verification.

Release gate: before the app leaves the preview, each release will publish a verification dataset with expected outputs on this page, so you can run your own checks.

Scientific computation runs in your browser, and individual apps need no account. Optional Bench accounts and backups use the site server and configured providers. The privacy statement, the hosts it contacts and the headers its workspace is served with are generated from its registry entry, the same record that sets the browser's policy.

What leaves your browser

Hosts the app contacts

• pubchem.ncbi.nlm.nih.gov

  Operator

  NCBI, U.S. National Library of Medicine

  What for

  Retrieves the selected compound's complete formula and structure through this site.

  What it receives

  the compound name searched or selected

  When

  on lookup

Headers the workspace is served with