def test_polymerize_functionCall(self):
sequences = np.array([[0, 1, 0, 1], [3, 1, 0, 2], [2, 2, 1, 1]])
baseAmounts = np.array([9, 9, 9, 9])
energy = 100
# Good calls test
polymerize(sequences, baseAmounts, energy, np.random.RandomState())
def test_polymerize_testVariableSequenceLength(self):
sequences = np.array([[0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 1, 1], [3] * 12,
[2] * 12, [1, 1, 1] + [P] * 9])
baseAmounts = np.array([30] * 4)
energy = 50
result = polymerize(sequences, baseAmounts, energy,
np.random.RandomState())
assert_equal(result.sequenceElongation, np.array([12, 12, 12, 3]))
def test_polymerize_testEnergyLimited(self):
sequences = np.array([[0, 1, 0, 1], [1, 1, 1, 1]])
baseAmounts = np.array([9, 9])
energy = 6
result = polymerize(sequences, baseAmounts, energy,
np.random.RandomState())
assert_equal(result.sequenceElongation, np.array([3, 3]))
assert_equal(result.monomerUsages, np.array([2, 4]))
self.assertEqual(6, result.nReactions)
def test_polymerize_testAllAvailableBases(self):
sequences = np.array([[0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 1, 1], [3] * 12,
[2] * 12, [P] * 12])
baseAmounts = np.array([11] * 4)
energy = 30
result = polymerize(sequences, baseAmounts, energy,
np.random.RandomState())
assert_equal(result.sequenceElongation, np.array([12, 9, 9, 0]))
assert_equal(result.monomerUsages, np.array([3, 5, 11, 11]))
self.assertEqual(30, result.nReactions)
def test_polymerize_testFairness(self):
sequences = np.array([[0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3], [3] * 12,
[2] * 12, [P] * 12])
baseAmounts = np.array([11] * 4)
baseAmountsOriginal = baseAmounts.copy()
energy = 30
result = polymerize(sequences, baseAmounts, energy,
np.random.RandomState())
assert_equal(result.sequenceElongation, np.array([10, 9, 9, 0]))
assert_equal(result.monomerUsages, np.array([3, 3, 11, 11]))
self.assertEqual(28, result.nReactions)
self.assertTrue((baseAmounts == baseAmountsOriginal).all(),
'should not modify monomerLimits')
def _fullProfile():
np.random.seed(0)
sequences, monomerLimits, reactionLimit, randomState = _setupRealExample()
# Recipe from https://docs.python.org/2/library/profile.html#module-cProfile
pr = cProfile.Profile()
pr.enable()
sequenceElongation, monomerUsages, nReactions = polymerize(
sequences, monomerLimits, reactionLimit, randomState)
pr.disable()
s = StringIO.StringIO()
sortby = 'cumulative'
ps = pstats.Stats(pr, stream=s).sort_stats(sortby)
ps.print_stats()
print s.getvalue()
def test_polymerize_testVariableElongation(self): sequences = np.array([ [0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 1, 1], [3] * 12, [2] * 12, [1, 1, 1] + [P] * 9]) baseAmounts = np.array([30] * 4) energy = 20 rates = np.array([4, 3, 2, 1]) result = polymerize( sequences, baseAmounts, energy, np.random.RandomState(), rates, variable_elongation=True) assert_equal(result.sequenceElongation, np.array([8, 6, 4, 2]))
def _simpleProfile():
np.random.seed(0)
sequences, monomerLimits, reactionLimit, randomState = _setupExample()
nSequences, length = sequences.shape
nMonomers = monomerLimits.size
sequenceLengths = (sequences != PAD_VALUE).sum(axis=1)
t = time.time()
sequenceElongation, monomerUsages, nReactions = polymerize(
sequences, monomerLimits, reactionLimit, randomState)
evalTime = time.time() - t
assert (sequenceElongation <= sequenceLengths + 1).all()
assert (monomerUsages <= monomerLimits).all()
assert nReactions <= reactionLimit
assert nReactions == monomerUsages.sum()
print """
Polymerize function report:
For {} sequences of {} different monomers elongating by at most {}:
{:0.1f} ms to evaluate
{} polymerization reactions
{:0.1f} average elongations per sequence
{:0.1%} monomer utilization
{:0.1%} energy utilization
{:0.1%} fully elongated
{:0.1%} completion
""".format(nSequences, nMonomers, length, evalTime * 1000, nReactions,
sequenceElongation.mean(),
monomerUsages.sum() / monomerLimits.sum(),
nReactions / reactionLimit,
(sequenceElongation == sequenceLengths).sum() / nSequences,
sequenceElongation.sum() / sequenceLengths.sum())
def evolveState(self):
## Module 1: Replication initiation
# Get number of active DNA polymerases and oriCs
activeDnaPoly = self.activeDnaPoly.molecules()
activePolymerasePresent = (len(activeDnaPoly) > 0)
oriCs = self.oriCs.molecules()
n_oric = len(oriCs)
n_chromosomes = self.full_chromosome.total()[0]
# If there are no chromosomes and oriC's, return immediately
if n_oric == 0 and n_chromosomes == 0:
return
# Get cell mass
cellMass = (self.readFromListener("Mass", "cellMass") * units.fg)
# Get critical initiation mass for simulation medium environment
current_nutrients = self._external_states['Environment'].nutrients
self.criticalInitiationMass = self.getDnaCriticalMass(
self.nutrientToDoublingTime[current_nutrients])
# Calculate mass per origin of replication, and compare to critical
# initiation mass. This is a rearrangement of the equation:
# (Cell Mass)/(Number of origins) > Critical mass
# If the above inequality holds true, initiate a round of chromosome
# replication for every origin of replication
massFactor = cellMass / self.criticalInitiationMass
massPerOrigin = massFactor / n_oric
replication_initiated = False
# If conditions are true, initiate a round of replication on every
# origin of replication
if massPerOrigin >= 1.0 and self.partialChromosomes.counts().sum(
) == 0:
replication_initiated = True
# Get replication round indexes of active DNA polymerases
if activePolymerasePresent:
replicationRound = activeDnaPoly.attr('replicationRound')
else:
# Set to -1 to set values for new polymerases to 0
replicationRound = np.array([-1])
# Get chromosome indexes of current oriCs
chromosomeIndexOriC = oriCs.attr('chromosomeIndex')
# Calculate number of new DNA polymerases required (4 per origin)
n_new_polymerase = 4 * n_oric
# Add new polymerases and oriC's
activeDnaPolyNew = self.activeDnaPoly.moleculesNew(
"dnaPolymerase", n_new_polymerase)
oriCsNew = self.oriCs.moleculesNew("originOfReplication", n_oric)
# Calculate and set attributes of newly created polymerases
sequenceIdx = np.tile(np.array([0, 1, 2, 3]), n_oric)
sequenceLength = np.zeros(n_new_polymerase, dtype=np.int)
replicationRound = np.ones(
n_new_polymerase, dtype=np.int) * (replicationRound.max() + 1)
# Polymerases inherit index of the OriC's they were initiated from
chromosomeIndexPolymerase = np.repeat(chromosomeIndexOriC, 4)
activeDnaPolyNew.attrIs(
sequenceIdx=sequenceIdx,
sequenceLength=sequenceLength,
replicationRound=replicationRound,
chromosomeIndex=chromosomeIndexPolymerase,
)
# Calculate and set attributes of newly created oriCs
# New OriC's share the index of the old OriC's they were
# replicated from
oriCsNew.attrIs(chromosomeIndex=chromosomeIndexOriC)
# Write data from this module to a listener
self.writeToListener("ReplicationData", "criticalMassPerOriC",
massPerOrigin)
self.writeToListener("ReplicationData", "criticalInitiationMass",
self.criticalInitiationMass.asNumber(units.fg))
## Module 2: replication elongation
# If no active polymerases are present, return immediately
# Note: the new DNA polymerases activated in the previous module are
# not elongated until the next timestep.
if len(activeDnaPoly) == 0:
return
# Build sequences to polymerize
dNtpCounts = self.dntps.counts()
sequenceIdx, sequenceLengths, massDiffDna = activeDnaPoly.attrs(
'sequenceIdx', 'sequenceLength', 'massDiff_DNA')
sequences = buildSequences(self.sequences, sequenceIdx,
sequenceLengths,
self._dnaPolymeraseElongationRate())
# Use polymerize algorithm to quickly calculate the number of
# elongations each "polymerase" catalyzes
reactionLimit = dNtpCounts.sum()
result = polymerize(sequences, dNtpCounts, reactionLimit,
self.randomState)
sequenceElongations = result.sequenceElongation
dNtpsUsed = result.monomerUsages
# Compute mass increase for each polymerizing chromosome
massIncreaseDna = computeMassIncrease(
sequences, sequenceElongations,
self.polymerized_dntp_weights.asNumber(units.fg))
updatedMass = massDiffDna + massIncreaseDna
# Update positions of each "polymerase"
updatedLengths = sequenceLengths + sequenceElongations
activeDnaPoly.attrIs(sequenceLength=updatedLengths,
massDiff_DNA=updatedMass)
# Update counts of polymerized metabolites
self.dntps.countsDec(dNtpsUsed)
self.ppi.countInc(dNtpsUsed.sum())
## Module 3: replication termination
# Determine if any polymerases reached the end of their sequences. If
# so, terminate replication and update the attributes of the remaining
# polymerases and OriC's to reflect the new chromosome structure.
terminalLengths = self.sequenceLengths[sequenceIdx]
didTerminate = (updatedLengths == terminalLengths)
terminatedChromosomes = np.bincount(sequenceIdx[didTerminate],
minlength=self.sequences.shape[0])
# If any of the polymerases were terminated, check if all polymerases
# initiated the same round as the terminated polymerases has already
# been removed - if they have, update attributes of the remaining
# polymerases and oriC's, and remove the polymerases. This is not done
# when some polymerases were initiated in the same timestep, as the
# attributes for these new polymerases cannot be updated in the same
# timestep.
if didTerminate.sum() > 0 and replication_initiated == False:
# Get attributes from active DNA polymerases and oriC's
sequenceIdx, chromosomeIndexPolymerase, replicationRound = activeDnaPoly.attrs(
'sequenceIdx', 'chromosomeIndex', 'replicationRound')
chromosomeIndexOriC = oriCs.attr('chromosomeIndex')
# Check that all terminated polymerases were initiated in the same
# replication round
assert np.unique(replicationRound[didTerminate]).size == 1
# Get chromosome indexes of the terminated polymerases
chromosomeIndexesTerminated = np.unique(
chromosomeIndexPolymerase[didTerminate])
newChromosomeIndex = chromosomeIndexPolymerase.max() + 1
# Get replication round index of the terminated polymerases
terminatedRound = replicationRound[didTerminate][0]
for chromosomeIndexTerminated in chromosomeIndexesTerminated:
# Get all remaining active polymerases initiated in the same
# replication round and in the given chromosome
replicationRoundMatch = (replicationRound == terminatedRound)
chromosomeMatch = (
chromosomeIndexPolymerase == chromosomeIndexTerminated)
remainingPolymerases = np.logical_and(replicationRoundMatch,
chromosomeMatch)
# Get all terminated polymerases in the given chromosome
terminatedPolymerases = np.logical_and(didTerminate,
chromosomeMatch)
# If all active polymerases are terminated polymerases, we are
# ready to split the chromosome and update the attributes.
if remainingPolymerases.sum() == terminatedPolymerases.sum():
# For each set of polymerases initiated in the same
# replication round, update the chromosome indexes to a new
# index for half of the polymerases.
for roundIdx in np.arange(terminatedRound + 1,
replicationRound.max() + 1):
replicationRoundMatch = (replicationRound == roundIdx)
polymerasesToSplit = np.logical_and(
replicationRoundMatch, chromosomeMatch)
n_matches = polymerasesToSplit.sum()
# Number of polymerases initiated in a single round
# must be a multiple of eight.
assert n_matches % 8 == 0
# Update the chromosome indexes for half of the polymerases
secondHalfIdx = np.where(polymerasesToSplit)[0][(
n_matches // 2):]
chromosomeIndexPolymerase[
secondHalfIdx] = newChromosomeIndex
# Reset chromosomeIndex for active DNA polymerases
activeDnaPoly.attrIs(
chromosomeIndex=chromosomeIndexPolymerase)
# Get oriC's in the chromosome getting divided
chromosomeMatchOriC = (
chromosomeIndexOriC == chromosomeIndexTerminated)
n_matches = chromosomeMatchOriC.sum()
# Number of OriC's in a dividing chromosome should be even
assert n_matches % 2 == 0
# Update the chromosome indexes for half of the OriC's
secondHalfIdx = np.where(chromosomeMatchOriC)[0][(
n_matches // 2):]
chromosomeIndexOriC[secondHalfIdx] = newChromosomeIndex
# Reset chromosomeIndex for oriC's
oriCs.attrIs(chromosomeIndex=chromosomeIndexOriC)
# Increment the new chromosome index in case another
# chromosome needs to be split
newChromosomeIndex += 1
# Delete terminated polymerases
activeDnaPoly.delByIndexes(np.where(didTerminate)[0])
# Update counts of newly created chromosome halves. These will be
# "stitched" together in the ChromosomeFormation process
self.partialChromosomes.countsInc(terminatedChromosomes)
def evolveState(self):
ntpCounts = self.ntps.counts()
self.writeToListener("GrowthLimits", "ntpAllocated", ntpCounts)
activeRnaPolys = self.activeRnaPolys.molecules()
if len(activeRnaPolys) == 0:
return
# Determine sequences that can be elongated
rnaIndexes, transcriptLengths, massDiffRna = activeRnaPolys.attrs('rnaIndex', 'transcriptLength', 'massDiff_mRNA')
sequences = buildSequences(self.rnaSequences, rnaIndexes, transcriptLengths, self.rnapElngRate)
ntpCountInSequence = np.bincount(sequences[sequences != polymerize.PAD_VALUE], minlength = 4)
# Polymerize transcripts based on sequences and available nucleotides
reactionLimit = ntpCounts.sum()
result = polymerize(sequences, ntpCounts, reactionLimit, self.randomState)
sequenceElongations = result.sequenceElongation
ntpsUsed = result.monomerUsages
nElongations = result.nReactions
# Calculate changes in mass associated with polymerization and update active polymerases
massIncreaseRna = computeMassIncrease(sequences, sequenceElongations, self.ntWeights)
updatedMass = massDiffRna + massIncreaseRna
didInitialize = (transcriptLengths == 0) & (sequenceElongations > 0)
updatedLengths = transcriptLengths + sequenceElongations
updatedMass[didInitialize] += self.endWeight
activeRnaPolys.attrIs(transcriptLength = updatedLengths, massDiff_mRNA = updatedMass)
# Determine if transcript has reached the end of the sequence
terminalLengths = self.rnaLengths[rnaIndexes]
didTerminate = (updatedLengths == terminalLengths)
terminatedRnas = np.bincount(rnaIndexes[didTerminate], minlength = self.rnaSequences.shape[0])
# Remove polymerases that have finished transcription from unique molecules
activeRnaPolys.delByIndexes(np.where(didTerminate)[0])
nTerminated = didTerminate.sum()
nInitialized = didInitialize.sum()
nElongations = ntpsUsed.sum()
# Update bulk molecule counts
self.ntps.countsDec(ntpsUsed)
self.bulkRnas.countsIs(terminatedRnas)
self.inactiveRnaPolys.countInc(nTerminated)
self.ppi.countInc(nElongations - nInitialized)
# Calculate stalls
expectedElongations = np.fmin(self.rnapElngRate, terminalLengths - transcriptLengths)
rnapStalls = expectedElongations - sequenceElongations
# Write outputs to listeners
self.writeToListener("TranscriptElongationListener", "countRnaSynthesized", terminatedRnas)
self.writeToListener("TranscriptElongationListener", "countNTPsUSed", nElongations)
self.writeToListener("GrowthLimits", "ntpUsed", ntpsUsed)
self.writeToListener("RnapData", "rnapStalls", rnapStalls)
self.writeToListener("RnapData", "ntpCountInSequence", ntpCountInSequence)
self.writeToListener("RnapData", "ntpCounts", ntpCounts)
self.writeToListener("RnapData", "expectedElongations", expectedElongations.sum())
self.writeToListener("RnapData", "actualElongations", sequenceElongations.sum())
self.writeToListener("RnapData", "didTerminate", didTerminate.sum())
self.writeToListener("RnapData", "terminationLoss", (terminalLengths - transcriptLengths)[didTerminate].sum())
def evolveState(self):
# Write allocation data to listener
self.writeToListener("GrowthLimits", "gtpAllocated", self.gtp.count())
self.writeToListener("GrowthLimits", "aaAllocated", self.aas.counts())
# Get number of active ribosomes
activeRibosomes = self.activeRibosomes.molecules()
self.writeToListener("GrowthLimits", "activeRibosomeAllocated",
len(activeRibosomes))
if len(activeRibosomes) == 0:
return
# Build amino acids sequences for each ribosome to polymerize
proteinIndexes, peptideLengths, massDiffProtein = activeRibosomes.attrs(
'proteinIndex', 'peptideLength', 'massDiff_protein')
sequences = buildSequences(self.proteinSequences, proteinIndexes,
peptideLengths, self.ribosomeElongationRate)
if sequences.size == 0:
return
# Calculate elongation resource capacity
aaCountInSequence = np.bincount(
sequences[(sequences != polymerize.PAD_VALUE)])
aaCounts = self.aas.counts()
# Using polymerization algorithm elongate each ribosome up to the limits
# of amino acids, sequence, and GTP
result = polymerize(
sequences,
aaCounts,
10000000, # Set to a large number, the limit is now taken care of in metabolism
self.randomState)
sequenceElongations = result.sequenceElongation
aasUsed = result.monomerUsages
nElongations = result.nReactions
# Update masses of ribosomes attached to polymerizing polypeptides
massIncreaseProtein = computeMassIncrease(sequences,
sequenceElongations,
self.aaWeightsIncorporated)
updatedLengths = peptideLengths + sequenceElongations
didInitialize = ((sequenceElongations > 0) & (peptideLengths == 0))
updatedMass = massDiffProtein + massIncreaseProtein
updatedMass[didInitialize] += self.endWeight
# Write current average elongation to listener
currElongRate = (sequenceElongations.sum() /
len(activeRibosomes)) / self.timeStepSec()
self.writeToListener("RibosomeData", "effectiveElongationRate",
currElongRate)
# Update active ribosomes, terminating if neccessary
activeRibosomes.attrIs(peptideLength=updatedLengths,
massDiff_protein=updatedMass)
# Ribosomes that reach the end of their sequences are terminated and
# dissociated into 30S and 50S subunits. The polypeptide that they are polymerizing
# is converted into a protein in BulkMolecules
terminalLengths = self.proteinLengths[proteinIndexes]
didTerminate = (updatedLengths == terminalLengths)
terminatedProteins = np.bincount(
proteinIndexes[didTerminate],
minlength=self.proteinSequences.shape[0])
activeRibosomes.delByIndexes(np.where(didTerminate)[0])
self.bulkMonomers.countsInc(terminatedProteins)
nTerminated = didTerminate.sum()
nInitialized = didInitialize.sum()
self.ribosome30S.countInc(nTerminated)
self.ribosome50S.countInc(nTerminated)
# Update counts of amino acids and water to reflect polymerization reactions
self.aas.countsDec(aasUsed)
self.h2o.countInc(nElongations - nInitialized)
# Report stalling information
expectedElongations = np.fmin(self.ribosomeElongationRate,
terminalLengths - peptideLengths)
ribosomeStalls = expectedElongations - sequenceElongations
# Write data to listeners
self.writeToListener("GrowthLimits", "aasUsed", aasUsed)
self.writeToListener("GrowthLimits", "gtpUsed", self.gtpUsed)
self.writeToListener("RibosomeData", "ribosomeStalls", ribosomeStalls)
self.writeToListener("RibosomeData", "aaCountInSequence",
aaCountInSequence)
self.writeToListener("RibosomeData", "aaCounts", aaCounts)
self.writeToListener("RibosomeData", "expectedElongations",
expectedElongations.sum())
self.writeToListener("RibosomeData", "actualElongations",
sequenceElongations.sum())
self.writeToListener(
"RibosomeData", "actualElongationHist",
np.histogram(sequenceElongations, bins=np.arange(0, 23))[0])
self.writeToListener(
"RibosomeData", "elongationsNonTerminatingHist",
np.histogram(sequenceElongations[~didTerminate],
bins=np.arange(0, 23))[0])
self.writeToListener("RibosomeData", "didTerminate",
didTerminate.sum())
self.writeToListener("RibosomeData", "terminationLoss",
(terminalLengths -
peptideLengths)[didTerminate].sum())
self.writeToListener("RibosomeData", "numTrpATerminated",
terminatedProteins[self.trpAIndex])
self.writeToListener("RibosomeData", "processElongationRate",
self.ribosomeElongationRate / self.timeStepSec())