🌙 Dark mode

Workflow Overview

uSort-M transforms a pooled DNA library into an arrayed collection of sequence-verified single clones in approximately one week.

uSort-M workflow diagram

Timeline

Day Step Duration Notes
1 Pooled assembly 2-3 hours Golden Gate or Gibson assembly
1 Transformation 1-2 hours Standard chemical transformation
1-2 Overnight growth 12-16 hours Recovery on selective plates
2+ FACS sorting ~8 min/plate Single-cell isolation into 384-well
2+ Outgrowth 12-16 hours Overnight in 384-well plates
3+ PCR barcoding ~50 min/plate Well-specific barcode addition
3+ Pooling 1-2 hours Combine barcoded amplicons
4-6 Sequencing 1-3 days Plasmidsaurus or ONT
6+ Analysis 1-2 hours Demux, variant calling, reporting
6+ Hit picking Variable Automated or manual cherry-picking

Step-by-Step

1. Library Design

Define your variant library as a list of sequences or mutations. uSort-M works with any pooled DNA source:

  • Oligo pools (e.g. Twist Oligo Pools, IDT oPools)
  • Error-prone PCR products
  • Multiplex assembly libraries

Estimate costs and coverage:

usortm estimate --library-size 500 --seq-length 300 --fold-sampling 4

This command calculates projected costs for synthesis, cloning, sorting, barcoding, sequencing, and hit-picking. It also estimates the timeline and effort required for your experiment.

Create an experiment plan:

usortm plan variants.csv --output my_project/ \
  --seq-length 300 \
  --fold-sampling 4

The plan command reads your variant CSV, calculates the number of plates needed, generates barcode assignments, and creates a project directory with a detailed experimental plan.

Learn more about library design →

2. Pooled Assembly

Perform pooled Golden Gate or Gibson assembly to insert your variant sequences into the destination vector.

Key considerations:

  • Use a high-fidelity assembly method
  • Include a selectable marker
  • Target 10-100x transformation coverage

3. FACS Sorting

Use fluorescence-activated cell sorting to isolate single cells into 384-well plates.

Typical parameters:

  • Single-cell mode
  • FSC/SSC gating for live cells
  • ~8 minutes per 384-well plate
  • Sort 8-10x library size for good coverage

Detailed sorting protocol →

4. PCR Barcoding

After overnight outgrowth, add well-specific DNA barcodes via PCR.

Recommended barcode sets:

  • LevSeq (Arnold lab) - Optimized for Oxford Nanopore
  • evSeq - Dual-indexing for Illumina and ONT

Barcoding protocol details →

5. Sequencing

Pool barcoded amplicons and submit for sequencing.

Recommended platforms:

  • Commercial Long-read Sequencing - Plasmidsaurus Custom Sequencing offers affordable multi-Gb library sequencing with fast turnaround

6. Demultiplexing & Analysis

Use the uSort-M CLI to process sequencing data:

Demultiplex sequencing reads:

# Using library CSV (recommended — auto-generates reference FASTA)
usortm demux project/ --fastq data.fastq --library-csv variants.csv

# Or with a pre-built reference FASTA
usortm demux project/ --fastq data.fastq --reference reference.fasta

The demux command runs the LevSeq barcode pipeline: reference alignment with minimap2, strand splitting, barcode demux with Dorado, per-well consensus generation, and variant calling.

Generate pick list:

usortm pick project/

Creates an Integra ASSIST PLUS-compatible CSV file for automated cherry-picking of sequence-verified clones, ordered to match your input library.

Generate analysis report:

usortm report project/

Produces reports (HTML, CSV, JSON) with coverage statistics, plate maps, and variant recovery metrics.

Demultiplexing details →

7. Hit Picking

Cherry-pick sequence-verified clones into a final arrayed format.

The usortm pick command generates Integra ASSIST PLUS-compatible hit lists for automated liquid handling.

File Formats

Variant CSV

Input file listing your library variants. Required columns:

name,sequence
variant_001,ATGGCTAAAGGTGAAGAACTG...
variant_002,ATGGCTAGAGGTGAAGAACTG...
variant_003,ATGGCTAAAGGTGAGGAACTG...

Alternative format using mutations (for substitution libraries):

name,mutation
variant_001,A123G
variant_002,T456C
variant_003,G789A

Barcode CSV

Well-specific DNA barcodes for demultiplexing. Format:

plate,well,barcode
plate_1,A01,ACGTACGT
plate_1,A02,TGCATGCA
plate_1,A03,GCTAGCTA

Recommended barcode sets: LevSeq (Arnold lab, ONT-optimized) or evSeq (dual-indexing for Illumina/ONT).

Reference FASTA

Expected sequence for variant calling:

>gene_of_interest
ATGGCTAAAGGTGAAGAACTGTTTACCGGTGTTGTGCCGATTCTGGTGGAACTGGATGGT
GATGTGAACGGTCACAAATTCTCTGTGCGTGGTGAAGGTGAAGGTGATGCTACCTACGGT

Expected Results

For a typical 500-variant library with 8x oversampling:

Metric Expected Value
Wells sorted ~4,000
Wells with growth ~2,700 (67%)
Unique variants recovered ~450 (90%)
Cost per variant ~$6-8
Total time 5-7 days