Back

Design Automation of Microfluidic Single and Double Emulsion Droplets with Machine Learning

Lashkaripour, A.; McIntyre, D. P.; Calhoun, S. G. K.; Krauth, K.; Densmore, D. M.; Fordyce, P. M.

2023-06-02 bioengineering
10.1101/2023.05.31.543128 bioRxiv
Show abstract

Droplet microfluidics enables kHz screening of picoliter samples at a fraction of the cost of other high-throughput approaches. However, generating stable droplets with desired characteristics typically requires labor-intensive empirical optimization of device designs and flow conditions that limit adoption to specialist labs. Here, we compile the most comprehensive droplet dataset to date and use it to train machine learning models capable of accurately predicting device geometries and flow conditions required to generate stable aqueous-in-oil and oil-in-aqueous single and double emulsions from 15 to 250 m at rates up to 12000 Hz for different fluids commonly used in life sciences. Novel device geometries predicted by our models for as-yet-unseen fluids yield accurate predictions, establishing their generalizability. Finally, we generate an easy-to-use design automation tool that yield droplets within 3 m (< 8%) of the desired diameter, facilitating tailored droplet-based platforms for new applications and accelerating their utility in life sciences.

Matching journals

The top 10 journals account for 50% of the predicted probability mass.

1
Advanced Science
○ 286 papers in training set
Top 0.3%
10.5%
Similar papers in this journal
2
Lab on a Chip
● 96 papers in training set
Top 0.2%
8.8%
Similar papers in this journal
3
Small
○ 78 papers in training set
Top 0.1%
7.2%
Similar papers in this journal
4
Advanced Materials Technologies
○ 29 papers in training set
Top 0.1%
5.1%
Similar papers in this journal
5
Nature Communications
○ 5641 papers in training set
Top 31%
4.3%
Similar papers in this journal
6
Science Advances
● 1243 papers in training set
Top 8%
4.0%
7
Analytical Chemistry
● 218 papers in training set
Top 0.9%
3.4%
8
Biofabrication
○ 36 papers in training set
Top 0.3%
2.8%
9
PLOS ONE
● 5266 papers in training set
Top 41%
2.4%
10
PLOS Computational Biology
● 1863 papers in training set
Top 12%
2.4%
50% of probability mass above
11
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 24%
2.1%
12
SLAS Technology
○ 14 papers in training set
Top 0.1%
2.1%
13
Bioengineering & Translational Medicine
○ 21 papers in training set
Top 0.2%
2.1%
14
Computational and Structural Biotechnology Journal
● 242 papers in training set
Top 3%
2.1%
15
Advanced Functional Materials
○ 46 papers in training set
Top 0.6%
2.1%
16
Scientific Reports
○ 3612 papers in training set
Top 49%
1.9%
17
Communications Chemistry
○ 48 papers in training set
Top 0.5%
1.7%
18
Small Methods
○ 29 papers in training set
Top 0.3%
1.7%
19
Advanced Healthcare Materials
○ 85 papers in training set
Top 1%
1.5%
20
eLife
● 5828 papers in training set
Top 56%
1.3%
21
ACS Biomaterials Science & Engineering
● 37 papers in training set
Top 0.7%
1.1%
22
Frontiers in Bioengineering and Biotechnology
○ 98 papers in training set
Top 2%
1.1%
23
Biomaterials
○ 84 papers in training set
Top 1%
1.1%
24
Cell Systems
○ 201 papers in training set
Top 4%
1.1%
25
Annals of Biomedical Engineering
○ 37 papers in training set
Top 0.9%
1.1%
26
Journal of Neural Engineering
○ 221 papers in training set
Top 2%
1.0%
27
Physics of Fluids
● 13 papers in training set
Top 0.1%
1.0%
28
Biotechnology and Bioengineering
○ 53 papers in training set
Top 0.8%
1.0%
29
Advanced Materials Interfaces
○ 10 papers in training set
Top 0.2%
0.8%
30
Biotechnology Journal
○ 11 papers in training set
Top 0.2%
0.8%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.