Release Notes

Fluxion

What's New

• The liver now takes up lactate at the measured rate, and the red cell's and muscle's glycolytic intermediates sit closer to measured values. Three equilibrium constants in the model were off from the measured values: lactate dehydrogenase's was four times too high for body temperature, enolase's about six times too low and phosphoglycerate mutase's about twice too high. Corrected, the liver takes up about 170 µmol of lactate per kg per minute at rest (it took about 45; measured about 150), the red cell's phosphoenolpyruvate-to-3-phosphoglycerate ratio rises from 0.08 to 0.34 (measured about 0.3), its fructose-1,6-bisphosphate and glyceraldehyde-3-phosphate fall into the measured range and its dihydroxyacetone phosphate and 3-phosphoglycerate fall about 4× and 1.7× (still above measured), and muscle's 3-phosphoglycerate falls from 70 to 12 µM. The shuttle capacities that set the muscle's and liver's cytoplasmic redox were re-centered to match, the blood set points are unchanged, every guided exercise teaches the same lesson and the Tissue health page reads the same. The trade-offs are listed under Known Issues.

• Muscle now holds its pH during sustained exercise. At a moderate workload a real muscle holds its pH for hours, but the model's turned steadily acidic, falling from 6.95 to 6.70 in 20 minutes while blood lactate stayed low, and the muscle's energy collapsed within half an hour. The cause was a bookkeeping error in the protons of the citric acid cycle. Pyruvate dehydrogenase, isocitrate dehydrogenase and 2-oxoglutarate dehydrogenase were each releasing a proton that their chemistry does not give, while citrate synthase, the carboxylases, glycogen synthase, glycerol kinase, PFK-2 and HMG-CoA synthase were each leaving out one they do. The cell released two protons too many for every pyruvate it burned. At rest that was hidden by terms tuned against it, but it grows with the rate of burning, so it was about a tenth of every ATP at exercise. All of these now balance charge exactly, and a new test checks every reaction. At a moderate workload (0.2 mM/s, about 17 times rest) the muscle's pH settles within minutes and stays near 7.0 for the hour, blood pH stays above 7.3 and blood lactate near 1 mM. At the hardest setting, which is above what a real muscle can sustain, the muscle's pH now falls steadily (7.1 to 6.8 over 15 minutes), blood lactate climbs to about 17 mM and blood pH to 7.18, and the muscle keeps its ATP rather than collapsing, as a real muscle does until it is exhausted. The resting body is unchanged. The strong-uncoupling exercise now settles at a membrane potential near 85 mV rather than 73 mV at ten minutes, and every guided exercise teaches the same lesson.

• Muscle now burns fat. Resting muscle used to run on sugar alone, which is not how a real muscle works. It now takes fatty acids from the blood and gets about half of its resting energy from fat, as measured in the human leg (a respiratory quotient of about 0.81). Its glucose uptake falls to the measured rate as a result. Fat burning switches off when insulin is high, standing in for the signal that blocks fat entry into mitochondria after a meal, and steps up as the muscle contracts. Resting blood glucose settles slightly higher (about 4.6 mM), because muscle now uses less of it. Every guided exercise teaches the same lesson. The model has no fat stored inside muscle yet, so working muscle burns less fat than a real one.

• A new Tissue health page shows which tissues are in trouble, and why. The Body view's new Tissue health button, in the top-left corner of the body diagram beside Tissue overview, opens a page that reads every tissue and the bloodstream live during a run and marks each as healthy, stressed, dying or dead (the bloodstream as healthy, stressed or critical). For each tissue in trouble it lists the metabolites responsible: the current value, the safe range, and whether it is too high or too low. Dead and dying tissues lead the page, and you don't have to open it to notice them: while any tissue is in trouble, a colored badge beside the button shows the worst status, such as "1 dead" or "2 stressed", and hovering over it names the tissues. The ranges are human clinical and physiological danger levels, such as the ATP loss at which cells die, the blood pH and glucose levels associated with high mortality, and the methemoglobin levels that cause coma, rather than the textbook normal ranges, so the resting body reads healthy everywhere. A tissue whose energy has collapsed reads dying, and is marked dead only once it has stayed collapsed as long as the real tissue survives that: about an hour for muscle and liver, and half a day for red cells. From then on it stays dead for the rest of the run, even if a lowered workload brings its chemistry back, because a real tissue would not recover. The oxygen exercise's Anaerobic setting cuts oxygen to the whole body, so the page shows the liver dying within a minute: it has the highest oxygen demand per kilogram of the modeled tissues and no phosphocreatine reserve.

• A new Tissue overview shows every tissue at a glance. Previously, comparing tissues meant switching the whole-body view's active organ one at a time. The Body view's new Tissue overview button now opens a page showing every registered tissue's own key readouts — muscle and liver's fuel, energy, and membrane potential; the red cell's redox and glutathione state — side by side, plus a blood ledger of the shared bloodstream's fuels, hormones, and gases in one place. It updates live during a run and is not tied to whichever organ you have open elsewhere.

• Every tissue now represents its full real-body mass. Skeletal muscle, the liver, and the red cells were previously sized against the shared bloodstream with hand-set weights — the red cell in particular was a small representative sample. Each is now sized from reference human anatomy (29 kg of muscle, 1.8 kg of liver, 2.3 L of red cells, against 14 L of extracellular fluid), so each tissue's share of the blood's glucose, lactate, oxygen, and fat budget reflects the real body. The body rests exactly where it did before, and every guided exercise teaches the same lesson; what changes is how strongly each organ moves the bloodstream during a transient.

• The liver now runs at its real metabolic rate. Its oxygen use, its resting energy demand, and the glucose it supplies to the body now match human measurements. The liver has the mitochondrial density of a real liver cell and its own resting energy demand, and the hormone switches a real liver uses: glucagon turns on glycogen breakdown, a glucokinase brake stops the fasting liver from re-trapping the glucose it releases, and fasting idles pyruvate dehydrogenase so pyruvate is saved for making new glucose. At rest the body is in the post-absorptive state its blood chemistry describes, so the liver now supplies the body's glucose itself — its measured share, about 0.66 mmol per minute, with the kidney supplying the rest — and blood glucose holds steady when dietary glucose stops instead of sliding toward hypoglycemia. Its glycogen store is sized to real livers. The resting bloodstream is unchanged and every guided exercise teaches the same lesson.

• Resting muscle now runs at its real metabolic rate, and the brain takes its own share of blood glucose. Resting muscle used to burn about three times the energy real resting muscle does, draw about ten times the glucose, and export lactate while at rest. It now matches human measurements of resting energy turnover, oxygen use, and glucose uptake, keeps its glycogen store steady at rest, and holds free ADP, phosphocreatine, and phosphate at measured resting levels. The changes are real muscle biology: the calcium released by contraction switches on glycogen breakdown, ATP production responds to ADP the way it does in living muscle, the citric acid cycle's dehydrogenases are held back by their products the way the purified enzymes are, and the glycerol-phosphate shuttle runs in its one real direction, which also corrects the redox balance of both the muscle and liver cytoplasm. Muscle glycogen starts at measured human levels. A new stand-in draws the brain's ~0.4 mmol of glucose per minute from the blood, so cutting dietary glucose no longer also switches the brain off. The resting bloodstream is unchanged and every guided exercise teaches the same lesson; the oxygen exercise's Hypoxic setting is now a clearer midpoint between aerobic and anaerobic, and each exercise's one-line interpretation was re-checked against the running model.

• The red blood cell now runs at its real metabolic rate. Red cells used to burn nearly four times the glucose real red cells do and never settled — sugar phosphates piled up and the cell's NAD⁺ was trapped as NADH. They now run on their measured enzyme activities and on the controls real red cells use: their 2,3-bisphosphoglycerate feeds back on hexokinase and phosphofructokinase, and their own pyruvate kinase isozyme. They take up about 1.2 µmol of glucose per mL of cells per hour, make about two lactates per glucose, and reach a true steady state. A long-standing error in one of glycolysis's equilibrium constants (glyceraldehyde-3-phosphate dehydrogenase, entered in the wrong concentration units) is also corrected; that fixes the redox balance of the red cell and the size of the triose-phosphate pools in muscle and liver. The favism exercise teaches the same lesson, and the pyruvate kinase exercise's interpretation now reads PEP relative to pyruvate kinase flux.

• The glossary and hover cards describe the model as it now runs. The glucose supply is described as what it has become, a small blood-side stand-in for the kidney, with the liver supplying the body's glucose at rest; glycogen synthase is described working through UDP-glucose; muscle phosphorylase is switched on by the calcium of contraction, and the liver's by glucagon; the red cell's 2,3-bisphosphoglycerate brakes on hexokinase and phosphofructokinase are listed; and blood glucose and 2,3-bisphosphoglycerate have their own entries. Blood protons are now described as the pool blood pH is read from. Every reaction and metabolite in the model now has its own glossary entry and hover card (74 were missing, including the liver's glucose-making and ketone pathways, the pentose phosphate pathway, the red cell's 2,3-bisphosphoglycerate shunt, the hormones and the blood pools).

• Simulations start at rest. Every run used to begin with the lower half of glycolysis empty, so its first stretch was a warm-up rather than a resting cell: cytosolic NADH spiked, the membrane potential drifted, and at the model's realistic resting rates it took about an hour to settle. Each tissue and the blood now start from the model's own resting state, built from the measured pool sizes, so a resting run is flat from the first second and every guided exercise responds to its preset from rest. The uncoupling exercise now names each preset correctly from its first seconds, and a solver error that could stop a run when the membrane potential passed through zero (the strong-uncoupling preset, or a run with the respiratory chain switched off) is fixed. Starting from the true resting state also showed that the liver's NADPH balance had been set on a run that had not yet settled; it now settles where it was meant to.

• Hard exercise is no longer limited by oxygen. The body's oxygen delivery could only rise about threefold above rest, so hard whole-body exercise drained the blood of oxygen and the liver, now running at its real energy demand, ran out of energy; and working muscle could draw oxygen from the blood only about 16 times faster than at rest. Delivery can now rise about 25-fold, roughly what a typical adult reaches at maximal effort, with blood oxygen staying near its resting level, and muscle can take up oxygen as fast as that delivery allows. The liver holds its normal energy state throughout, and the hardest exercise setting no longer runs the blood or the liver out of oxygen. Resting values and every guided exercise's lesson are unchanged.

Known Issues

• Sustained exercise still runs the muscle slowly down. A muscle holds its pH at moderate work, but the model's cannot hold the work for ever: its mitochondria supply about 84% of the ATP and glycolysis and phosphocreatine the rest, limited by how fast fuel reaches the citric acid cycle (muscle has no contraction-stimulated glucose uptake and no fat stored inside it), so over an hour at the moderate setting phosphocreatine falls to a few millimolar, ATP drifts down by about 5%, and pH begins to slip after the first 40 minutes. At the hardest setting the muscle sustains the work for the 15-minute run, with pH falling steadily and lactate climbing. The workload you set is held until you change it; to model a muscle tiring or resting, add a scheduled change that lowers its ATP demand.

• The model has no fed state yet. The resting body is post-absorptive, and there is not yet a way to give it a meal: the glucose-supply control only nudges blood glucose toward 5 mM, so a postprandial rise in glucose and insulin, and the liver's switch to storing glucose, cannot be shown.

• The liver still draws on its glycogen faster than a real liver at rest. It now takes up lactate at the measured rate, but oxidizes most of it instead of turning it into glucose, and makes several times the ketones it did (an artificial disposal removes them), so about 88% of its glucose still comes from glycogen and the store would last hours rather than a day.

• The red cell's resting pyruvate is too high. With the measured enolase and mutase constants its glycolytic intermediates fall toward measured levels, but pyruvate rests near 0.6 mM (measured 0.05–0.08) and its lactate-to-pyruvate ratio near 0.7 (measured about 10), because its pyruvate kinase's constants were set for the old, much higher, intermediate levels. Its glucose use, ATP and the favism exercise are unchanged.

• Some resting muscle values sit outside measured ranges. Resting muscle's cytoplasm runs more oxidized than real muscle (NAD⁺/NADH about 2,500 against about 700), which leaves its lactate-to-pyruvate ratio low (about 4 against 10–20), puts its glycerol-3-phosphate about 2.4 times above the measured level, and pushes the malate-aspartate shuttle's amino acids to one side: glutamate rests near zero while aspartate and oxaloacetate rest high. Because runs now start at rest, these show up directly as the starting values.

• The muscle NADPH-backup exercise is good for its 20 minutes. With muscle at its real resting metabolic rate, a G6PD-deficient muscle under the exercise's heavy oxidant challenge keeps its glutathione for the full exercise but gives out after about 25–30 minutes if a run is extended: the backup enzymes use up a cytosolic pool that real muscle refills through pathways the model does not yet include.