InDels analysis of CRISPR lines by NGS amplicon sequencing technology for a multicopy gene family.

In this work, we present a workflow to analyze InDels from the multicopy α-gliadin gene family from wheat based on NGS data without the need to previously establish a reference sequence for each genetic background. The pipeline was tested it in a multiple sample set, including three generations of edited wheat lines (T0, T1, and T2), from three different backgrounds and ploidy levels (hexaploid and tetraploid). Implementation of Bayesian optimization of Usearch parameters, inhouse Python, and bash scripts are reported.

Bayesian optimization was implemented to optimize Usearch v9.2.64 parameters from merge to search steps for the α-gliadin amplicons on the wild type lines.

python Step1_Bayesian_usearch.py --database <str> --file_intervals <str> --trim_primers <str> --path_usearch_control <str>

Help:

Outputs:

• Bayesian_usearch.txt File with optimal values, optimal function value, samples or observations, obatained values and search space.
• Bayesian.png Convergence plot.
• Bayesian_data_res.txt File with the minf(x) after n calls in each iteration.

Usearch pipeline optimazed on wild type lines for studying results of optimization.

Step2_usearch_WT_to_DB.sh dif pct maxee amp id path_control name_dir_usearch path_database trim_primers

Help:

Arguments must be disposed in the order indicated before.

• dif Optimal value for dif Usearch parameter.
• pct Optimal value for pct Usearch parameter.
• maxee Optimal value for maxee Usearch parameter.
• amp Optimal value for amp Usearch parameter.
• id Optimal value for id Usearch parameter.
• path_control Path of the wild type lines fastq files.
• name_dir_usearch Path of Usearch.
• path_database Path of alpha-gliadin amplicon database.
• trim_primers Trim primers in reads if you use database without primers. Optios: YES | NO.

Outputs:

Usearch merge files, filter files, unique amplicons file, unique denoised amplicon (Amp/otu) file, otu table file.

Usearch pipeline optimazed on all lines (wild types and CRISPR lines) for studying denoised unique amplicon relative abundances.

Step3_usearch_ALL_LINES.sh dif pct maxee amp id path_ALL name_dir_usearch trim_primers

Help:

Arguments must be disposed in the order indicated before.

• dif Optimal value for dif Usearch parameter.
• pct Optimal value for pct Usearch parameter.
• maxee Optimal value for maxee Usearch parameter.
• amp Optimal value for amp Usearch parameter.
• id Optimal value for id Usearch parameter.
• path_ALL Path of all lines (wild type and CRISPR lines) fastq files.
• name_dir_usearch Path of Usearch.
• trim_primers Trim primers in reads if you use database without primers. Optios: YES | NO.

Outputs:

Usearch merge files, filter files, unique amplicon file, unique denoised amplicon (Amp/otu) file, otu table file.

Before Step 4, otu table file must be normalized by TMM normalization method (edgeR package in R). Results of TMM normalized unique denoised amplicons table can be represented as heatmaps. Unique denoised amplicons can be compared between them to detect Insertions and Deletions (InDels) in CRISPR lines.

Create tables with the presence or absence of unique denoised amplicons in each CRISPR line compared to the wild type lines.

python Step4_usearch_to_table.py --file_otu <str> --file_group <str> --prefix_output <str> --genotype <str>

Help:

Default threshold 0.3 % of frequency of each unique denoised amplicon (Amp) in each line.

Outputs:

Substitute "name" in output names for the prefix_output string.

• Amptable_frequency.txt Table of Amps (otus) transformed to frequencies for apply the threshold.
• Amptable_brutes_name.txt Table with number of reads contained in the unique denoised amplicons (Amps) present in each line.
• Amps_name.txt Table with number of unique denoised amplicons (Amps) in each line.

Python 3.6 or later is required.