Biology: 3D anatomy, physiology and lab tools
3D anatomy for the whole body, physiology, genetics and ecology simulators from the action potential to an SIR epidemic, and clinical and lab calculators.
75 tools
The largest part of this section is 3D anatomy: explorers built on the BodyParts3D dataset, from the teeth to the whole body, with every structure named in the dataset’s own words. Alongside them are physiology simulators, where the action potential, the nephron and the oxygen dissociation curve run as models you can change, an SIR model of how an epidemic grows and then stops, and a predator and prey pair whose populations cycle rather than settle, and calculators for the bench and the clinic.
Genetics is the group that runs on counting rather than measuring, and counting is where the judgement is. A monohybrid cross predicts three to one, so 80 offspring should split 60 to 20, and a real count of 65 to 15 looks like evidence of something until a chi-square test puts it at 1.67 on one degree of freedom, which chance alone exceeds in about one cross in five. The same arithmetic applied to a population rather than a cross is the Hardy-Weinberg relation: a recessive disorder seen in one birth in 10,000 makes about one person in 50 a carrier, which is why carriers, not sufferers, hold nearly all of a rare allele.
The sequence tools all read from one copy of the standard genetic code, NCBI translation table 1, so no two of them can disagree about what a codon means. That is worth saying because the usual failure in this corner is not arithmetic: it is a frame. Drop one base from a coding sequence and every codon after it is read in a different frame, and what comes out is a plausible-looking protein that is wrong from that point on.
Molecular biology runs on a small number of conversions that appear in nearly every protocol: absorbance to concentration, RPM to relative centrifugal force, two cell counts to a doubling time. None are difficult, and all of them are easy to get wrong at the end of a long day.
The calculators for them spell out the assumption behind each one, because the assumption is usually where the error lives. The extinction factor applies to DNA but not RNA. The rotor radius is what turns a speed into a force. The growth equation only holds while growth is exponential.
Each of these numbers is a proxy: absorbance for mass of nucleic acid, rotor speed for the force on a pellet, two counts for a growth rate. A proxy is fine until its assumption fails, and the calculation does not stop when it does: it returns a figure that still looks reasonable. Knowing where each breaks is worth more than knowing the formula.
3D anatomy
The 3D anatomy hub arranges these by part of the body and by system.
Simulators
Calculators
Visualisers and practice
Absorbance measures everything at once
A reading at 260 nm counts every nucleotide the beam passes through: your intact target, degraded fragments, free nucleotides left over from the extraction, and any RNA sitting in a DNA prep. It cannot distinguish them, so it reports a total rather than a quantity of the thing you care about.
The extinction factor you then apply assumes you already know what the sample is. An absorbance of 1.0 across a 1 cm path is roughly 50 µg/mL for double-stranded DNA, 40 for RNA, and 33 for single-stranded DNA and short oligonucleotides. Applying the double-stranded factor of 50 to an RNA sample overstates the concentration by about 25 percent, and nothing in the result flags it.
That is why the purity ratios matter. A 260/280 ratio near 1.8 for DNA or 2.0 for RNA suggests the absorbance is genuinely coming from nucleic acid, a lower figure means protein or phenol is contributing, and a 260/230 ratio below about 2.0 points to guanidine, EDTA or carbohydrate. Keep the reading between 0.1 and 1.0: below that the signal sits in the instrument’s noise, above it most spectrophotometers stop being linear. When the answer has to be right, a dye-based fluorometric assay only responds to what it binds.
Always quote spins in × g, never RPM
RPM describes how fast the rotor turns, not how hard the sample is pushed. Relative centrifugal force scales with the radius, so the same 10,000 RPM in a large rotor delivers substantially more force than in a small one. A protocol written in RPM is only reproducible on the exact centrifuge it was written for.
Convert to × g and the instruction travels. The one number you need is the rotor radius, from its documentation. A typical microcentrifuge rotor of about 8.5 cm turns 13,000 RPM into roughly 16,000 × g, which is where that familiar pair of figures comes from. Use r-max for pelleting, since that is where material collects, and note that guessing the radius from the tube length typically costs 20 to 50 percent. Speed matters more still, because force goes as its square: a 10 percent error in RPM is a 21 percent error in force, which is often the difference between a tight pellet and a smear.
A growth rate only means something in exponential phase
Doubling time is defined for exponential growth and nothing else. If your two counts straddle a lag phase, or the culture has hit confluence or exhausted its medium, the equation still returns a figure, but it averages across a period that was not exponential and will overstate the true doubling time.
Choose the interval so it spans several doublings, because the counts enter through a logarithm. If the two counts differ by a factor of 2, a 10 percent error in one shifts the doubling time by about 14 percent; if they differ by a factor of 16, the same error shifts it by about 3 percent. Waiting three or four doublings between samples buys more accuracy than counting either sample more carefully.
It helps to know what normal looks like. E. coli in rich medium manages about 20 minutes, yeast about 90 minutes, HeLa cells roughly 24 hours, and primary human fibroblasts 30 to 40 hours. Counting is noisier than it feels, too: cells in a chamber follow Poisson statistics, so a count of 100 carries about 10 percent relative error before any pipetting is added.
Mass, moles and why fragment length changes the answer
One pair of units is identical and routinely treated as though it were not: 1 ng/µL is exactly 1 µg/mL. Both appear in protocols, and a reading reported in one against a target written in the other is a needless source of doubt.
Mass and molar amount differ, and the conversion depends on length. A base pair of double-stranded DNA averages about 650 g/mol, so a 1 kb fragment is around 650,000 g/mol and 1 µg of it is roughly 1.5 pmol. Because molecule count scales inversely with length, 100 ng of a 200 bp fragment holds about 50 times as many molecules as 100 ng of a 10 kb fragment. Equal masses are not equal numbers of molecules, so a reaction set up by mass across different fragment lengths is not the reaction you designed.
Common questions
Why do my absorbance and fluorometric DNA readings disagree?
Because they measure different things. An A260 reading counts every nucleotide in the sample, including degraded fragments, free nucleotides and any RNA carried through the extraction, while a dye-based fluorometric assay responds only to what the dye binds, typically intact double-stranded DNA. The absorbance figure is therefore usually the higher of the two, and the size of the gap is itself informative: a large discrepancy points to RNA or degradation rather than an instrument fault. Where the input amount matters, such as library preparation, trust the fluorometric number.
How do I find my rotor radius if I do not have the documentation?
Look up the rotor model number, which is usually engraved on the rotor body or printed inside the lid, and take the r-max figure from the manufacturer’s specification. Measuring from the centre of the spindle to the bottom of a seated tube is a workable fallback, but estimating from the tube length alone typically introduces a 20 to 50 percent error in the resulting force. Record the radius alongside the protocol once you have it, so future spins can be quoted in × g and reproduced on any machine.