noise2read 0.1.2

Turn noise to read

Turn ‘noise’ to signal: accurately rectify millions of erroneous short reads through graph learning on edit distances

noise2read, originated in a computable rule translated from PCR erring mechanism that: a rare read is erroneous if it has a neighboring read of high abundance, turns erroneous reads into their original state without bringing up any non-existing sequences into the short read set(<300bp) including DNA and RNA sequencing (DNA/RNA-seq), small RNA, unique molecular identifiers (UMI) and amplicon sequencing data.

Click noise2read to jump to its documentation

Quick-run example

Quick-run example for testing noise2read by setting only 1 trial for Optuna and 10 estimators for xGboost which are not the parameters used in our paper.

• noise2read installation

Please refer to QuickStart or Installation.

• Clone the codes with datasets in github

git clone https://github.com/Jappy0/noise2read
cd noise2read/Examples/simulated_miRNAs

• Quick-run testing noise2read on D14

  • with high ambiguous errors correction and using GPU for training (running about 4 mins with 26 cores and GPU)

  noise2read -m correction -c ../../config/Quick_test.ini -a True -g gpu_hist

Examples for correcting simulated miRNAs data with mimic UMIs by noise2read

Take data sets D14 and D16 as examples.

• noise2read installation

Please refer to QuickStart or Installation.

• Clone the codes with datasets in github

git clone https://github.com/Jappy0/noise2read
cd noise2read/Examples/simulated_miRNAs

• Reproduce the evaluation results for D14 and D16 from raw, true and corrected datasets

noise2read -m evaluation -i ./simulated_miRNAs/raw/D14_umi_miRNA_mix.fa -t ./simulated_miRNAs/true/D14_umi_miRNA_mix.fa -r ./simulated_miRNAs/correct/D14_umi_miRNA_mix.fasta -d ./result
noise2read -m evaluation -i ./simulated_miRNAs/raw/D16_umi_miRNA_subs.fa -t ./simulated_miRNAs/true/D16_umi_miRNA_subs.fa -r ./simulated_miRNAs/correct/D16_umi_miRNA_subs.fasta -d ./result

• correcting D14

  • with high ambiguous errors correction and using GPU for training

  noise2read -m correction -c ../../config/D14.ini -a True -g gpu_hist

  • without high ambiguous errors correction and using CPU (default) for training

  noise2read -m correction -c ../../config/D14.ini -a False

  Note

  Noise2read may produce slightly different corrected result from these under Examples/simulated_miRNAs/correct

• correcting D16

  • with high ambiguous errors correction and using GPU for training

  noise2read -m correction -c ../../config/D16.ini -a True -g gpu_hist

  • without high ambiguous errors correction and using CPU (default) for training

  noise2read -m correction -c ../../config/D16.ini -a False

  Note

  Noise2read may produce slightly different corrected result from these under Examples/simulated_miRNAs/correct

Examples for correcting outcome sequence of ABEs and CBEs by noise2read

• Clone the codes

git clone https://github.com/Jappy0/noise2read
cd noise2read/CaseStudies
mkdir ABEs_CBEs
cd ABEs_CBEs

• Download datasets D32_D33.

• Using noise2read to correct the datasets. The running time of each experiment is about 13 minutes using 26 cores and GPU for training.

noise2read -m correction -i ./D32_D33/raw/D32_ABE_outcome_seqs.fasta -a False -d ./ABE/
noise2read -m correction -i ./D32_D33/raw/D33_CBE_outcome_seqs.fasta -a False -d ./CBE/