Amplicon Indel Hunter (AIH) is a large (>5-bp) indel detection method that is reference genome independent and highly sensitive for the identification of somatic indels in amplicon-based, paired-end, NGS data.
C++ Tcl HTML Makefile C Python Other
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
aiHunter_example_data
binary_distribution
License.txt
README.md
aiHunter.py
aiHunter_Documentation.pdf
amplicon_indel_diagnoser.py
amplicon_indel_hunter.py
convert_aid_to_vcf.py

README.md

Amplicon Indel Hunter (AIH/AID) - Indel detection in amplicon-based, paired-end, NGS data

aiHunter v1.1.0

Contact

Sabah Kadri: skadri@bsd.uchicago.edu

Jeremy P Segal: jsegal@bsd.uchicago.edu

Publication (please cite if using software):

Sabah Kadri, Chao J. Zhen, Michelle N. Wurst, Bradley C. Long, Zi-Feng Jiang, Y. Lynn Wang, Larissa V. Furtado, Jeremy P. Segal, Amplicon Indel Hunter Is a Novel Bioinformatics Tool to Detect Large Somatic Insertion/Deletion Mutations in Amplicon-Based Next-Generation Sequencing Data, The Journal of Molecular Diagnostics, Volume 17, Issue 6, November 2015, Pages 635-643, ISSN 1525-1578, http://dx.doi.org/10.1016/j.jmoldx.2015.06.005.\newline (http://www.sciencedirect.com/science/article/pii/S1525157815001506)

Introduction

Accurate detection of large insertions and deletions (indels) via amplicon-based targeted NGS assays remains a challenge when depending on alignment-based methods. Sequencing reads that cover these indels are, by definition, different from the reference sequence, and lead to variable performance of alignment algorithms. Amplicon Indel Hunter (AIH) is a large (>5-bp) indel detection method that is reference genome independent and highly sensitive for the identification of somatic indels in amplicon-based, paired-end, NGS data. The software (aiHunter) takes as input paired end fastq files and information about the amplicons in the assay, and detects amplicons with potential large indel. The output from AIH is given as input to a helper tool, Amplicon Indel Diagnoser (AID) which tries to annotate the exact indel sequence and returns output in VCF format.

Release Notes

v1.1.0 has the following changes over v1.0.0:

  1. Bug fix in VCF file generation:
    If AID fails to decode the exact indel, due to bad quality or the size of the indel, it returns a
    Failed indel identification - manual review required for <amplicon name>
    line in the <read1filename>.R2fastq.finalindelstats file.
    In v1.0.0, the VCF file generation was failing if there was a decoding failure by AID. This is now fixed. The VCF file will only contain entries for the indels which are successfully decoded by AID. However, we encourage all users to also investigate the <read1 filename>.R2fastq.finalindelstats file to manually check entries which failed with AID.
  2. Example data:
    Example data, and its associated documentation is now included in the aiHunter folder. Please see the folder aiHunter_example_data.
  3. New FAQ section We have added a new FAQ section to this documentation, based on common issues faced by some of our users. Please refer to the FAQ section for better understanding of any errors encounters. Please feel free to reach out Sabah Kadri (skadri@bsd.uchicago.edu) for any additional trouble- shooting.

Quickstart

A quick way to start using aiHunter is to have paired end Fastq files, an infosheet with details of the assay, and a fasta file with the insert sequences. Please see "Input Data" below for more details.

python aiHunter.py --read1 <Read1 Fastq> --read2 <Read2 Fastq> --amp <Amplicon information sheet> --inserts <Inserts Fasta file> [options]

Installation

aiHunter has been tested with python v2.7.6. It requires the following packages in the python PATH to run it: (1) argparse (2) os (3) sys (4) time (5) Bio (biopython) (6) math
An executable binary is also provided with its own documentation in the binary_distribution folder. This can only be run on the Linux OS.

License

See License.txt

Input Data

aiHunter requires the following input files in order to run:

--read1 Read1 Fastq file.
--read2 Read2 Fastq file.
--amp Tab-delimited amplicon info file. See below.
--inserts Fasta file with insert sequences. See below.

Amplicon Info File

Please make sure that the coordinates listed in this file are accurate. These will be used by the Amplicon Indel Diagnoser program to identify the genomic coordinates of the indel. Note: Please make sure that the file has a header line. Tab delimited file with following columns:

Column name in header Description
AmpliconName Name of the amplicon. Depends on assay.  
Primer1Sequence Primer1 sequence. Should be + strand genomic sequence.
Primer2Sequence Primer2 sequence. Should be "-" strand genomic sequence, that is, reverse complement of the sense sequence.
AmpliconLength Length of the amplicon. Can be calculated as (AmpliconStop - AmpliconStart + 1) or (Primer2End - Primer1Start + 1)
Chr Reference Chromosome
GenomicAmpliconStart Genomic Amplicon Start, which is the same as Primer1 Start
GenomicAmpliconEnd Genomic Amplicon End, which is the same as Primer2 End

Inserts Fasta file

Amplicon Indel Diagnoser part of aiHunter requires a fasta file with the insert sequences of the assay in order to determine the exact sequence of the indel, wherever possible. This file can be generated using by making a BED file of the inserts and then using UCSC genome browser in order to download the Fasta file.

Note: Please make sure your inserts bed file has Chr Primer1End and (Primer2Start - 1) in order to get the correct sequence from UCSC

Note: Please make sure that the amplicon names in the inserts fasta file match the amplicon names in the amplicon info file

Optional parameters

--help or -h Outputs a quick help manual for running aiHunter.
--version or -v Provides version information.
--info Quick information about the Amplicon Info file
--out OUTDIR Output directory for the results. Default = Current working directory.
--cushion CUSHION Cushion Value (default: 5). Please see Section 3 or publication for more details.
--maf SIGNIFICANCE Mutant allele frequency threshold (default: 0.05)

Outputs from AIH and AID

Output files and formats

aiHunter outputs the following four files for each pair of fastq files. The main output file is <read1 filename>.R2fastq.finalindelstats.vcf :

  • <read1 filename>.R2fastq.indelcalls.txt

This is the output of the AIH module of the software. It is a table for one row for each amplicon. The four columns and their descriptions are given below.

Amplicon Name of amplicon as provided in the input info file
Reads_w_indel>5bp Number of Reads with potential indel of size greater than the cushion size. The 5bp represents the default cushion size, and even if a different cushion is provided, this column header remains the same.
#readpairspassingfilter Number of reads which match the primer pair for the specific amplicon with minimum identity of 90%
%indel Column#2 / Column#3
  • <read1 filename>.R2fastq.indelcalls.significant.txt

A list of amplicon names that pass the MAF threshold (See --maf SIGNIFICANCE). Each of these are investigated by the AID module of the software.

  • <read1 filename>.R2fastq.finalindelstats

Output of the AID module of the software. The file has no header. It is a 8-column tab-delimited text file with the following columns.
chrom position depth depth ref_allele ref_freq indel indel_freq

  • <read1 filename>.R2fastq.finalindelstats.vcf

Output of the AID module of the software. A VCF formatted output of the annotated (if possible) indels.

Indel Annotation Failure

In cases where the insertions are so large that there is little or no longer overlap between the read mates, AID might fail in determining the exact indel sequence and output:

Failed indel identification - manual review required for <amplicon name>

In such cases, we usually use a combination of the detailed logs from AID and inspection of the misaligned read alignments in IGV to manually determine the indel sequence. This is not a frequent occurrence and only affects insertions.

Bad quality Problems

Please note a bad sequencing run with deteriorating base qualities in a large portion of the read may affect the sensitivity of detection by this method, depending on the amplicon and read lengths.

FAQs

Q. The finalindelstats file returns Failed indel identification - manual review required for <amplicon name>

A. This can happen when the insertion is so large that there is little or no overlap between the read mates. In this case you might want to do a manual review. In such cases, we usually use a combination of the detailed logs from AID and inspection of the misaligned read alignments in IGV to manually determine the indel sequence.
This might also be due to a bad sequencing run with deteriorating base qualities in a large portion of the read may affect the sensitivity of detection by this method, depending on the amplicon and read lengths.

Q. The indelcalls.txt shows 0 counts for all amplicons.

A. There might be a problem with the input amplicon info file. Please recheck the format of the file, making sure that the Primer 1 sequence is in the genomic sense "+" direction and Primer 2 sequence is in the anti-sense "-" direction, that is, the reverse complement of the genomic sequence.
Hint: You can use UCSC Insilico PCR tool to check the orientations of the primer sequences

Q. The log file returns an error of "No sequence found for <amplicon name> and the finalindelstats file is empty even though the indelcalls.txt has high MAF for the amplicon.

A. Please make sure the insert sequences for each amplicon is present in the Fasta file, and that the names of the amplicons are the same between the fasta file and the amplicon info file.

Q. I am getting various string-related errors when I try to run the program.

A. Please make sure:

  1. The inserts.fa file has the Inserts - sequence of the amplicon between the primers and does not include the primer sequences.
  2. Sequences for each amplicon are present in the Fasta file, and that the names of the amplicons are the same between the fasta file and the amplicon info file.
  3. The Fastq files are not zipped. The current version of aiHunter only works with unzipped fastq files.
  4. All reads are of the same length, that is, the adapters are not trimmed off, either by the software on the instrument or other data pre-processing.