Dataset DOI: 10.5061/dryad.qnk98sfv8

This repository contains:

• Structural model of transmembrane helix dimers (PDB files)
• Sort-seq Data
• NGS data processing script (with example input data files)

Structural model of transmembrane helix dimers (PDB files)

Description of the data and file structure

The files include the CATM and AlphaFold3 (AF3) models of the twelve GAS-right dimers validated by the mutagenesis. The files also include the CATM and AF3 models of the three proteins whose mutagenesis matches an AF3 model but not a CATM model. The CATM and AF3 models have been aligned with each other based on the transmembrane helices.

Name scheme:

<Uniprot Code><protein name abbreviation><modeling program><model number><.pdb>

Files of proteins matching the CATM model:

• O75056_syndecan-3_AF3_model_0.pdb
• O75056_syndecan-3_CATM_model_1.pdb
• P02786_TfR1_AF3_model_4.pdb
• P02786_TfR1_CATM_model_1.pdb
• Q13591_sema-5A_AF3_model_1.pdb
• Q13591_sema-5A_CATM_model_1.pdb
• Q6P7N7_TMEM81_AF3_model_0.pdb
• Q6P7N7_TMEM81_CATM_model_1.pdb
• Q6UW88_Epigen_AF3_model_3.pdb
• Q6UW88_Epigen_CATM_model_1.pdb
• Q6UXE8_BTNL3_AF3_model_4.pdb
• Q6UXE8_BTNL3_CATM_model_3.pdb
• Q8N6P7_IL22R1_AF3_model_0.pdb
• Q8N6P7_IL22R1_CATM_model_1.pdb
• Q8NEA5_C19orf18_AF3_model_0.pdb
• Q8NEA5_C19orf18_CATM_model_2.pdb
• Q8NFY4_sema-6D_AF3_model_2.pdb
• Q8NFY4_sema-6D_CATM_model_1.pdb
• Q8TDQ0_HAVCR2_AF3_model_3.pdb
• Q8TDQ0_HAVCR2_CATM_model_1.pdb
• Q9H3T3_sema-6B_AF3_model_4.pdb
• Q9H3T3_sema-6B_CATM_model_1.pdb
• Q9NVM1_EVA1B_AF3_model_2.pdb
• Q9NVM1_EVA1B_CATM_model_1.pdb

Files of proteins matching only the AF3 model:

• P16410_CTLA4_AF3_model_3.pdb
• P16410_CTLA4_CATM_model_1.pdb
• P78380_OLR1_AF3_model_2.pdb
• P78380_OLR1_CATM_model_1.pdb
• Q9ULG6_CCPG1_AF3_model_0.pdb
• Q9ULG6_CCPG1_CATM_model_1.pdb

Code/software

PDB files are viewable with PyMol or other software that can read Protein Data Bank coordinate files.

Sort-seq Data

This repository contains processed data from high-throughput TOXGREEN sort-seq experiments used in the analysis presented in our manuscript. The data include deep sequencing results from three length libraries of single-pass membrane protein variants (wild-types and point mutants), as well as data describing the GFP fluorescence gates used during fluorescence-activated cell sorting.

The following files contain processed sequencing data for three libraries of different transmembrane domain lengths:

• allSeqs17.csv
• allSeqs19.csv
• allSeqs21.csv

Each row in these files corresponds to a unique sequence observed in a specific library, replicate, and fluorescence bin.

Column Descriptions:

Sort Gate Information File

• SortGateInfo.csv

This file contains data from the FACS gating and binning used during the sorting of each library replicate.

Column Descriptions:

NGS data processing script

This script fastqToTxt.pl processes NGS data in .fastq format by extracting high-quality reads, translating DNA sequences into protein sequences, and matching them against a reference list. It outputs a count and frequency of each detected sequence and preserves annotation information if the sequence is found in a reference sequence file.

Features

• Reads .fastq files and filters out low-quality reads.
• Extracts regions between given forward and reverse primers.
• Translates the DNA region into an amino acid sequence.
• Filters sequences by expected N- and C-terminal residues.
• Matches protein sequences to a reference sequence file.
• Reports the count, percentage, and metadata for matched sequences.

Input Files

1. FASTQ File (--seqFile)

A standard .fastq file with 4-line entries (sequence on line 2, quality on line 4).

2. Reference File (--refFile)

A tab-delimited file with 7 columns, where:

The protein sequence in column 6 is the key used to identify matches, and columns 0–5 are printed as annotation when a match is found.

Usage

perl fastqToTxt.pl --refFile reference.csv --seqFile reads.fastq --direction 1

Arguments

• --refFile: Path to the tab-delimited reference file.
• --seqFile: Path to the .fastq file to be analyzed.
• --direction: Direction of sequencing reads:
  • 1 = forward
  • 2 = reverse

Output

The script prints to standard output a tab-delimited table containing:

<Protein Sequence> <Count> <Fraction of total good reads> <Reference Annotation>

If the sequence is found in the reference file, the annotation from columns 0–5 is printed. If not, the script checks for special fallback cases or prints Unknown.

Notes

• Only sequences with expected quality (≤1 expected error) are retained.
• Primers are hard-coded and must match the sequences used in library construction.
• Sequences must start with "AS" and end with "L" after translation and are trimmed accordingly.
• Matches are based on exact protein sequences; partial or fuzzy matches are not considered.
• A default count threshold of 10 is applied to reduce noise.

Example input files

Run the script with the example input files provided, example_reference_file.csv and example.fastq with:

perl fastqToTxt.pl --refFile example_reference_file.csv --seqFile example.fastq --direction 1

Notes

• Sequence data files reflect processed reads that passed quality and frequency filters.
• Bin statistics in SortGateInfo.csv were used to reconstruct fluorescence values for each construct.
• All fluorescence values are reported in arbitrary units based on the GFP signal detected during sorting.

Contact

Alessandro Senes, senes@wisc.edu

Access information

Other publicly accessible locations of the data:

• The CATM files are available at catm.biochem.wisc.edu/CATM
• The AlphaFold files can be computed at the AlphaFold server https://alphafoldserver.com/

Data was derived from the following sources:

• The structural predictions were made from the protein sequence from the Uniprot database: https://www.uniprot.org/