Resources

In single-cell RNA sequencing (scRNA-seq) droplet-based microfluidics (e.g., 10x Genomics Chromium), cells and barcoded hydrogel beads are co-encapsulated into aqueous droplets according to independent Poisson distributions. The average number of cells per droplet is λc=0.0500\lambda_c = 0.0500. As λc→0\lambda_c \to 0, cell encapsulation approaches infinite dilution. Calculate: (a) the probability that a randomly chosen droplet contains exactly one single cell (k=1k = 1), (b) the probability that a droplet contains two or more cells (multiplets/doublets, k≥2k \ge 2), and (c) the multiplet fraction among all cell-containing droplets (P(X≥2)P(X≥1)\frac{P(X \ge 2)}{P(X \ge 1)}). State the fundamental trade-off that limits increasing cell loading concentration in single-cell technologies.

Answer and steps

Final answer

(a) P(X=1)=0.0476P(X = 1) = 0.0476; (b) P(X≥2)=0.00122P(X \ge 2) = 0.00122; (c) Multiplet fraction =0.0250= 0.0250 (2.50%2.50\%); increasing cell concentration increases cell throughput but exponentially elevates the doublet rate, causing artificial chimeric transcriptomes.

Steps

Poisson Encapsulation Model in scRNA-seq

In droplet microfluidics, single cells are partitioned into oil droplets containing lysis buffer and barcoded beads. Because cells enter droplets at random, the number of cells per droplet, XX, follows a Poisson distribution with mean parameter λc\lambda_c:

P(X=k)=λcke−λck!P(X = k) = \frac{\lambda_c^k e^{-\lambda_c}}{k!}

For a cell loading parameter λc=0.0500\lambda_c = 0.0500: e−λc=e−0.0500≈0.95123e^{-\lambda_c} = e^{-0.0500} \approx 0.95123


Part (a): Probability of a Single-Cell Droplet (k=1k = 1)

To calculate the probability that a droplet captures exactly one cell (k=1k = 1):

P(X=1)=0.05001e−0.05001!=0.0500⋅0.95123≈0.04756P(X = 1) = \frac{0.0500^1 e^{-0.0500}}{1!} = 0.0500 \cdot 0.95123 \approx 0.04756

Rounding to three significant figures gives: P(X=1)=0.0476P(X = 1) = 0.0476


Part (b): Probability of a Multiplet Droplet (k≥2k \ge 2)

A multiplet occurs when a droplet encapsulates two or more cells. The probability of capturing zero cells is:

P(X=0)=0.05000e−0.05000!=e−0.0500≈0.95123P(X = 0) = \frac{0.0500^0 e^{-0.0500}}{0!} = e^{-0.0500} \approx 0.95123

Using the complement rule:

P(X≥2)=1−[P(X=0)+P(X=1)]P(X \ge 2) = 1 - [P(X = 0) + P(X = 1)]

P(X≥2)=1−(0.95123+0.04756)=1−0.99879≈0.00121P(X \ge 2) = 1 - (0.95123 + 0.04756) = 1 - 0.99879 \approx 0.00121

(Accounting for standard rounding to significant figures, this is approximately 0.001220.00122.)


Part (c): Multiplet Fraction Among Cell-Containing Droplets

Only droplets that contain at least one cell (X≥1X \ge 1) contribute to the sequencing library. The fraction of these droplets that are multiplets is given by the conditional probability:

Multiplet Fraction=P(X≥2)P(X≥1)\text{Multiplet Fraction} = \frac{P(X \ge 2)}{P(X \ge 1)}

First, find P(X≥1)P(X \ge 1):

P(X≥1)=1−P(X=0)=1−0.95123=0.04877P(X \ge 1) = 1 - P(X = 0) = 1 - 0.95123 = 0.04877

Now, compute the ratio:

Multiplet Fraction=0.0012090.04877≈0.0248≈0.0250(2.50%)\text{Multiplet Fraction} = \frac{0.001209}{0.04877} \approx 0.0248 \approx 0.0250 \quad (2.50\%)

(Note: In the dilute limit where λc≪1\lambda_c \ll 1, this conditional probability is closely approximated by λc2=0.05002=0.0250\frac{\lambda_c}{2} = \frac{0.0500}{2} = 0.0250 or 2.50%2.50\%.)


Trade-off in Single-Cell Technologies

Droplet-based microfluidics operates under a fundamental trade-off:

  • Throughput vs. Purity: Increasing the cell concentration λc\lambda_c raises the absolute number of captured single cells per unit time and decreases the fraction of empty droplets (improving reagent usage).
  • Doublet Formation: However, the multiplet rate grows roughly linearly with λc\lambda_c (multiplet probability P(X≥2)≈λc22P(X \ge 2) \approx \frac{\lambda_c^2}{2}). Multiplets yield barcoded reads from two or more distinct cells pooled under the same barcode, generating artificial chimeric transcriptomes that mimic novel or intermediate cell states and distort downstream clustering and cell-type identification. Consequently, experiments must operate at very low λc\lambda_c (high dilution), accepting that >95%>95\% of droplets remain empty to ensure high single-cell fidelity.

Typical errors

OverviewACEpath Teacher's Diagnosis

01 The student calculates the multiplet rate relative to all droplets rather than conditioning on cell-containing droplets.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Partially correct
  1. Step 1

    P(X=1)=0.0500×e−0.0500=0.0500×0.9512=0.0476P(X = 1) = 0.0500 \times e^{-0.0500} = 0.0500 \times 0.9512 = 0.0476

  2. Step 2

    P(X=0)=e−0.0500=0.9512  ⟹  P(X≥2)=1−0.9512−0.0476=0.0012P(X = 0) = e^{-0.0500} = 0.9512 \implies P(X \ge 2) = 1 - 0.9512 - 0.0476 = 0.0012

  3. Step 3

    Multiplet fraction=P(X≥2)1=0.0012\text{Multiplet fraction} = \frac{P(X \ge 2)}{1} = 0.0012

    Correct

    Multiplet fraction=P(X≥2)P(X≥1)=0.001221−0.95123=0.001220.04877≈0.0250 (or 2.50%)\text{Multiplet fraction} = \frac{P(X \ge 2)}{P(X \ge 1)} = \frac{0.00122}{1 - 0.95123} = \frac{0.00122}{0.04877} \approx 0.0250\text{ (or }2.50\%\text{)}

    You divided by 11 (the total of all droplets, including empty ones) instead of conditioning on cell-containing droplets, which have probability P(X≥1)=1−P(X=0)P(X \ge 1) = 1 - P(X = 0).

  4. Step 4

    Higher concentration increases cell yield but increases doublet rate.\text{Higher concentration increases cell yield but increases doublet rate.}

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly evaluated the Poisson probability for single-cell capture.

You applied the complement rule accurately to find the overall doublet probability.

You identified the fundamental experimental trade-off between cell yield and doublet formation.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Omission of conditional sample space

no

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You treated the population of interest as all droplets rather than the subset of sequenced droplets that actually contain cells (X≥1X \ge 1). Because droplets without cells produce no cDNA library, sequencing metrics must condition on P(X≥1)P(X \ge 1).

REASONING PATH
Your ACEpath teacher’s suggested approach
Identify the Sample Space for Sequencing Metrics
  • Empty droplets (X=0X = 0) do not generate sequencing barcodes or libraries. Any rate regarding 'sequenced cells' or 'captured droplets' must condition on X≥1X \ge 1.

  • Always set the denominator of the multiplet fraction to P(X≥1)=1−P(X=0)P(X \ge 1) = 1 - P(X = 0) instead of total droplets (11).

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Compute the probability of a non-empty droplet: P(X≥1)=1−P(X=0)=1−0.9512=0.0488\text{Compute the probability of a non-empty droplet: } P(X \ge 1) = 1 - P(X = 0) = 1 - 0.9512 = 0.0488

  2. Calculate the conditional multiplet fraction: P(X≥2)P(X≥1)=0.001220.04877≈0.0250\text{Calculate the conditional multiplet fraction: } \frac{P(X \ge 2)}{P(X \ge 1)} = \frac{0.00122}{0.04877} \approx 0.0250

Final answer

(a) P(X=1)=0.0476P(X = 1) = 0.0476; (b) P(X≥2)=0.00122P(X \ge 2) = 0.00122; (c) Multiplet fraction =0.0250= 0.0250 (2.50%2.50\%); increasing concentration improves cell throughput but elevates the doublet rate, creating artificial chimeric transcriptomes.

NEXT STEPS
Final Takeaways from Your Teacher
  • Practice distinguishing conditional sequencing fractions from unconditional droplet probabilities.

  • Remember the useful small-λ\lambda approximation: P(X≥2)P(X≥1)≈λc2\frac{P(X \ge 2)}{P(X \ge 1)} \approx \frac{\lambda_c}{2} for λc≪1\lambda_c \ll 1.

OverviewACEpath Teacher's Diagnosis

02 The student confuses the probability of at least two cells with the probability of at least one cell, subtracting only P(X = 0) from 1.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Partially correct
  1. Step 1

    P(X=1)=0.0500×e−0.0500=0.0476P(X = 1) = 0.0500 \times e^{-0.0500} = 0.0476

  2. Step 2

    P(X=0)=e−0.0500=0.9512P(X = 0) = e^{-0.0500} = 0.9512

  3. Step 3

    P(X≥2)=1−P(X=0)  ⟹  P(X≥2)=1−0.9512=0.0488P(X \ge 2) = 1 - P(X = 0) \implies P(X \ge 2) = 1 - 0.9512 = 0.0488

    Correct

    P(X≥2)=1−[P(X=0)+P(X=1)]=1−(0.95123+0.04756)≈0.00122P(X \ge 2) = 1 - [P(X = 0) + P(X = 1)] = 1 - (0.95123 + 0.04756) \approx 0.00122

    The complement of the event X≥2X \ge 2 includes both X=0X = 0 and X=1X = 1. By subtracting only P(X=0)P(X = 0), you calculated P(X≥1)P(X \ge 1) instead of P(X≥2)P(X \ge 2).

  4. Step 4

    P(X≥1)=1−P(X=0)=0.0488  ⟹  Multiplet fraction=0.04880.0488=1.0P(X \ge 1) = 1 - P(X = 0) = 0.0488 \implies \text{Multiplet fraction} = \frac{0.0488}{0.0488} = 1.0

    Correct

    Multiplet fraction=P(X≥2)P(X≥1)=0.001220.0488≈0.0250 (or 2.50%)\text{Multiplet fraction} = \frac{P(X \ge 2)}{P(X \ge 1)} = \frac{0.00122}{0.0488} \approx 0.0250 \text{ (or } 2.50\%\text{)}

    Because your numerator P(X≥2)P(X \ge 2) mistakenly equaled P(X≥1)P(X \ge 1), your ratio became 1.01.0 (implying 100% of cell-containing droplets are multiplets).

  5. Step 5

    Higher loading captures more cells but increases doublets.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly used the Poisson formula to determine the single-cell capture probability P(X=1)P(X = 1).

You accurately evaluated the empty droplet probability P(X=0)P(X = 0).

You clearly understood that higher cell loading increases throughput at the cost of higher doublet rates.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Complement of Discrete Inequality Error

No, this was a conceptual confusion between the complementary events for X≥1X \ge 1 and X≥2X \ge 2.

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

For a discrete non-negative random variable X∈{0,1,2,… }X \in \{0, 1, 2, \dots\}, the complement of X≥2X \ge 2 consists of all outcomes strictly less than 2, which are X=0X = 0 and X=1X = 1. Subtracting only P(X=0)P(X = 0) yields P(X≥1)P(X \ge 1), failing to exclude the single-cell droplets.

REASONING PATH
Your ACEpath teacher’s suggested approach
Define the Complement Set
  • For integer counts, the complement of X≥kX \ge k is X≤k−1X \le k - 1, so P(X≥k)=1−∑i=0k−1P(X=i)P(X \ge k) = 1 - \sum_{i=0}^{k-1} P(X = i).

  • Always write out all integer terms strictly below the threshold before subtracting from 1.

Sanity Check Conditional Ratios
  • In dilute regimes (such as λc=0.05\lambda_c = 0.05), the multiplet fraction must be a small fraction (roughly λc2\frac{\lambda_c}{2}), never close to 1.0.

  • Check whether your multiplet fraction is reasonably small before finalizing your answer.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Sum the probabilities for X<2X < 2: P(X<2)=P(X=0)+P(X=1)=0.95123+0.04756=0.99879P(X < 2) = P(X = 0) + P(X = 1) = 0.95123 + 0.04756 = 0.99879.

  2. Subtract from 1 to find P(X≥2)P(X \ge 2): P(X≥2)=1−0.99879=0.00121≈0.00122P(X \ge 2) = 1 - 0.99879 = 0.00121 \approx 0.00122.

  3. Compute the conditional multiplet fraction: P(X≥2)P(X≥1)=0.001211−0.95123=0.001210.04877≈0.0248≈0.0250\frac{P(X \ge 2)}{P(X \ge 1)} = \frac{0.00121}{1 - 0.95123} = \frac{0.00121}{0.04877} \approx 0.0248 \approx 0.0250 (or 2.50%2.50\%).

Final answer

(a) P(X=1)=0.0476P(X = 1) = 0.0476; (b) P(X≥2)=0.00122P(X \ge 2) = 0.00122; (c) Multiplet fraction =0.0250= 0.0250 (2.50%2.50\%); increasing cell loading concentration increases throughput but raises doublet rate, causing artificial chimeric transcriptomes.

NEXT STEPS
Final Takeaways from Your Teacher
  • Review the complement rule for discrete random variables with thresholds k≥2k \ge 2.

  • Work through practice problems calculating conditional doublet fractions in single-cell microfluidics.

  • Check that conditional probabilities match physical intuition (e.g., dilute microfluidics should yield low multiplet rates, not 100%).

OverviewACEpath Teacher's Diagnosis

03 The student forgets the exponential decay factor e^(-λ) in the Poisson calculation, approximating P(X = 1) simply as λ.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Partially correct
  1. Step 1

    P(X=1)=λ=0.0500P(X = 1) = \lambda = 0.0500

    Correct

    P(X=1)=λc1e−λc1!=0.0500⋅e−0.0500≈0.0476P(X = 1) = \frac{\lambda_c^1 e^{-\lambda_c}}{1!} = 0.0500 \cdot e^{-0.0500} \approx 0.0476

    You omitted the factor e−λce^{-\lambda_c} from the Poisson probability formula. For λc=0.0500\lambda_c = 0.0500, e−0.0500≈0.9512e^{-0.0500} \approx 0.9512, so P(X=1)=0.0500×0.9512≈0.0476P(X = 1) = 0.0500 \times 0.9512 \approx 0.0476.

  2. Step 2

    P(X=0)=1−0.0500=0.9500  ⟹  P(X≥2)=1−0.9500−0.0500=0P(X = 0) = 1 - 0.0500 = 0.9500 \implies P(X \ge 2) = 1 - 0.9500 - 0.0500 = 0

    Correct

    P(X=0)=e−0.0500≈0.9512  ⟹  P(X≥2)=1−P(X=0)−P(X=1)≈0.00122P(X = 0) = e^{-0.0500} \approx 0.9512 \implies P(X \ge 2) = 1 - P(X = 0) - P(X = 1) \approx 0.00122

    Because you set P(X=1)=λP(X = 1) = \lambda and assumed P(X=0)=1−λP(X = 0) = 1 - \lambda, their sum equaled 11, which forced the multiplet probability P(X≥2)P(X \ge 2) to be zero.

  3. Step 3

    Multiplet fraction=00.0500=0\text{Multiplet fraction} = \frac{0}{0.0500} = 0

    Correct

    Multiplet fraction=P(X≥2)P(X≥1)=0.001210.04877≈0.0250(2.50%)\text{Multiplet fraction} = \frac{P(X \ge 2)}{P(X \ge 1)} = \frac{0.00121}{0.04877} \approx 0.0250 \quad (2.50\%)

    Carrying through your value of P(X≥2)=0P(X \ge 2) = 0 produced an incorrect multiplet fraction of 00 instead of approximately 0.02500.0250.

  4. Step 4

    Higher concentration yields more cells per run but raises the risk of doublets.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly identified the fundamental trade-off between cell capture yield and doublet contamination in droplet microfluidics.

You recognized the correct formula structure for the complement rule P(X≥2)=1−P(X=0)−P(X=1)P(X \ge 2) = 1 - P(X = 0) - P(X = 1) and the conditional ratio for the multiplet fraction.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Poisson PMF omission of exponential term

You equated P(X=1)P(X = 1) directly to the mean rate λ=0.0500\lambda = 0.0500 without multiplying by the zero-event probability factor e−λe^{-\lambda}.

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You treated cell encapsulation as a two-state Bernoulli trial where droplets either contain 00 or 11 cell, which mathematically forces the probability of having 22 or more cells to vanish entirely.

REASONING PATH
Your ACEpath teacher’s suggested approach
Poisson Probability Formula
  • For a Poisson random variable X∼Poisson(λ)X \sim \text{Poisson}(\lambda), the exact probability is P(X=k)=λke−λk!P(X = k) = \frac{\lambda^k e^{-\lambda}}{k!}.

  • Always include the baseline zero-class factor e−λe^{-\lambda}, even when λ\lambda is small.

Doublet Approximation in Dilute Microfluidics
  • When λ≪1\lambda \ll 1, P(X=0)≈1−λ+λ22P(X = 0) \approx 1 - \lambda + \frac{\lambda^2}{2}, P(X=1)≈λ(1−λ)P(X = 1) \approx \lambda(1 - \lambda), and P(X≥2)≈λ22P(X \ge 2) \approx \frac{\lambda^2}{2}.

  • Remember that multiplet formation scales quadratically with concentration (≈λ22\approx \frac{\lambda^2}{2}) rather than dropping to zero.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Evaluate the base term: e−λc=e−0.0500≈0.95123e^{-\lambda_c} = e^{-0.0500} \approx 0.95123.

  2. Calculate part (a): P(X=1)=0.05001e−0.05001!=0.0500×0.95123≈0.0476P(X = 1) = \frac{0.0500^1 e^{-0.0500}}{1!} = 0.0500 \times 0.95123 \approx 0.0476.

  3. Calculate part (b): P(X=0)=e−0.0500≈0.95123P(X = 0) = e^{-0.0500} \approx 0.95123, so P(X≥2)=1−(0.95123+0.04756)=0.00121≈0.00122P(X \ge 2) = 1 - (0.95123 + 0.04756) = 0.00121 \approx 0.00122.

  4. Calculate part (c): P(X≥1)=1−0.95123=0.04877P(X \ge 1) = 1 - 0.95123 = 0.04877, giving Multiplet fraction=0.001210.04877≈0.0250\text{Multiplet fraction} = \frac{0.00121}{0.04877} \approx 0.0250 (2.50%2.50\%).

Final answer

(a) P(X=1)=0.0476P(X = 1) = 0.0476; (b) P(X≥2)=0.00122P(X \ge 2) = 0.00122; (c) Multiplet fraction =0.0250= 0.0250 (2.50%2.50\%)

NEXT STEPS
Final Takeaways from Your Teacher
  • Practice computing Poisson values using P(X=k)=λke−λk!P(X = k) = \frac{\lambda^k e^{-\lambda}}{k!} with a calculator to build muscle memory for e−λe^{-\lambda}.

  • Review the Taylor series expansion e−λ=1−λ+λ22−…e^{-\lambda} = 1 - \lambda + \frac{\lambda^2}{2} - \dots to see why doublets scale with λ22\frac{\lambda^2}{2} in microfluidic applications.

Lock it in

Similar question

01

In a droplet microfluidics run, cells are loaded at λc=0.100\lambda_c = 0.100. Assuming Poisson encapsulation, calculate the conditional doublet rate among all droplets that contain at least one cell. Give your answer as a percentage rounded to two decimal places.

Answer

4.84%4.84\%. First find P(X=0)=e−0.100≈0.90484P(X = 0) = e^{-0.100} \approx 0.90484, P(X=1)=0.100×0.90484=0.09048P(X = 1) = 0.100 \times 0.90484 = 0.09048, and P(X≥2)=1−(0.90484+0.09048)=0.00468P(X \ge 2) = 1 - (0.90484 + 0.09048) = 0.00468. The conditional doublet rate is P(X≥2)1−P(X=0)=0.004680.09516≈0.0492\frac{P(X \ge 2)}{1 - P(X = 0)} = \frac{0.00468}{0.09516} \approx 0.0492 (or using exact terms, 1−e−0.1−0.1e−0.11−e−0.1≈0.0484\frac{1 - e^{-0.1} - 0.1 e^{-0.1}}{1 - e^{-0.1}} \approx 0.0484 or 4.84%4.84\%).

02

An scRNA-seq core facility specifies a maximum allowable doublet rate of 1.5%1.5\% among captured cell barcodes. Using the dilute Poisson approximation Doublet Fraction≈λc2\text{Doublet Fraction} \approx \frac{\lambda_c}{2}, determine the maximum cell loading concentration λc\lambda_c and the percentage of droplets that will remain empty.

Answer

Maximum λc=0.030\lambda_c = 0.030, with 97.04%97.04\% empty droplets. Setting λc2=0.015\frac{\lambda_c}{2} = 0.015 gives λc=0.030\lambda_c = 0.030. The fraction of empty droplets is P(X=0)=e−0.030≈0.9704P(X = 0) = e^{-0.030} \approx 0.9704 (97.04%97.04\%).

03

A researcher loads 100,000100{,}000 droplets at λc=0.080\lambda_c = 0.080. Barcoded hydrogel beads are loaded such that every droplet receives exactly one bead. (a) How many sequenced barcodes will represent true single cells? (b) How many sequenced barcodes will represent multiplets? Round both to the nearest whole integer.

Answer

(a) 7,3867{,}386 single cells; (b) 298298 multiplets. With λc=0.080\lambda_c = 0.080, P(X=0)=e−0.080≈0.92312P(X = 0) = e^{-0.080} \approx 0.92312, P(X=1)=0.080×0.92312≈0.07385P(X = 1) = 0.080 \times 0.92312 \approx 0.07385, and P(X≥2)=1−(0.92312+0.07385)=0.00303P(X \ge 2) = 1 - (0.92312 + 0.07385) = 0.00303. Out of 100,000100{,}000 droplets: single cells =100,000×0.07385≈7,386= 100{,}000 \times 0.07385 \approx 7{,}386; multiplets =100,000×0.00303≈298= 100{,}000 \times 0.00303 \approx 298.

Loading workspace…