Software tools for processing sequencing data from transposon insertion sequencing (TIS) experiments
Transposon insertion sequencing (TIS) harnesses the power and throughput of next generation sequencing to allow for genome-wide analysis of microbial gene function and pathways. Several methods for performing TIS library preparation and sequencing have been described (TnSeq, INSeq, HITS, TraDIS) and software tools such as INSeq_analysis and ESSENTIALS have been created to analyze sequencing results. TIS_tools is a set of perl scripts developed as a bridge between sequencing and analyisis by performing read alignments, filtering alignments, and outputting read count data in a format suitable for downstream analysis.
Please choose the approriate script based on which library prep method you used:
- "Goodman" protocol: INSeq_read_preprocess.pl
- "Stacey-Boll" protocol: INSeq_read_preprocess_Boll_protocol.pl
For sequence reads generated using the "INSeq" transposon insertion sequencing protocol as described by Goodman, Wu, and Gordon, performs the following processing steps:
- Separate multiplexed libraries by "barcode" sequence
- Identify and trim transposon sequence from reads
- Align reads to reference sequence(s)
- Assign aligments to positions in reference sequence
- Read count correction and filtering
- Output read counts
Prerequisites:
bowtie (version 0.12.8 or higher)
Usage:
perl INSeq_read_preprocess.pl [options] -r <reads.fastq> -c <barcodes.txt> -g <reference_sequence.fasta>
Required arguments:
-r Read sequence file, in fastq format. Can remain gzipped.
-c Tab-delimited text file of barcodes in the following format, one barcode per line:
pool_id<tab>barcode
Optional: If you are re-processing already de-barcoded and transposon-trimmed reads or if reprocessing previously-performed bowtie alignments, you can also include paths to either the processed read output files with or without paths to the bowtie alignment files. To include the sorted reads file (as previously output by this script): pool_id<tab>barcode<tab>/path/to/reads.fasta. To also include the bowtie alignment (as previously output by this program): pool_id<tab>barcode<tab>/path/to/reads.fasta<tab>/path/to/alignment.bowtie.
-g Reference genome sequence file, in fasta format. For best results, this file must include sequences for ALL genetic material present in the organism (i.e. chromosomes and plasmids).
Encouraged, but optional:
-w Sequence of actual parent strain that was used to produce transposon mutagenesis library, in fasta format. OK if this is a multi-contig draft assembly.
Other options:
-t Transposon sequence (default: "ACAGGTTG")
-m Number of transposon sequence mismatches allowed (default: 1)
-o Output format for count files. Possible options include:
- 'wiggl': Files will be in wiggle format that can be used as input to ESSENTIALS (default)
- 'inseq': Files will be in in the format of *_bowtiemap_processed.txt files produced by Goodman et al's INSeq_analysis software
- 'essen': Files will be in the format of allta_split_*.counts.txt files produced by ESSENTIALS
-s Minimum number of total reads (left flank + right flank) at an insertion site required for output (default: 3)
-d Minimum difference in read counts betwen sides to trigger investigation of site flanks. A higher number may result in faster processing, but lower sensitivity for identifying read count discrepancies. Minimum value is 1. (default: 1)
-h Ignore "shifted" reads, i.e. reads that align perfectly without mismatches, but one base upstream or downstream from a "TA" insertion site. (default: "shifted" reads will be added to a flank's total read count if addition of the reads results in the two flanks having a more balanced read count)
-b Path to directory containing bowtie and bowtie-build. For example: /Users/myname/applications/bowtie_folder (default: assumes this directory is in your PATH)
-p Number of parallel threads to run (default: 1)
Program output:
As the script runs, several statistics will be output to the screen. If you wish to capture these results, simply redirect STDOUT to a file like so:
perl INSeq_read_preprocess.pl -r reads.fastq -c barcodes.txt -g reference_sequence.fasta > stats.txt
or:
perl INSeq_read_preprocess.pl -r reads.fastq -c barcodes.txt -g reference_sequence.fasta | tee stats.txt
Barcode sorting, trimming, and transposon trimming results will take the following format:
Total reads: xxx
Total reads with barcode: xxx
Total reads with transposon: xxx
Transposons with 0 mismatch(es): xxx
Transposons with 1 mismatch(es): xxx
pool bc #_w_bc %_w_bc #_w_tn %_w_tn
...
Total reads is the total number of reads found in the input read fastq file
Total reads with barcode is the total number of reads found to have a recognizable barcode sequence at the start of the read sequence
Total reads with transposon is the total number of reads (from those found to have a barcode sequence) that had the given transposon sequence at the expected position (16 or 17 bp from the 5' end after removal of barcode sequence)
pool: Pool ID
bc: Barcode
#_w_bc: Number of reads with this barcode
%_w_bc: Percentage of total reads with this barcode
#_w_tn: Number of reads with this barcode found to have the given transposon sequence in the expected position
%_w_tn: Percentage of reads with this barcode found to have the given transposon sequence in the expected position
Output files:
Files beginning with "temp_" can be safely deleted.
Files will start with a prefix corresponding to the pool_id and barcode given in the barcode text file above. For example, if one of the pool IDs was "Input_1" identified with barcode "TATA", then all files relating to this pool will be prefixed with "Input_1_TATA".
<prefix>.reads.fasta: Fasta formatted file of all sorted, de-barcoded, and transposon trimmed read sequences associated with this pool.<prefix>.bowtie: Alignment file produced by bowtie<prefix>.<reference_sequence>.wiggle: Read alignment counts for each of the individual records that was present in the reference genome sequence file. For example, if the file given to the-roption contained one chromosomal sequence record and one plasmid sequence record there should be two of these files per barcode, each named according to the sequence record name and with coordinates corresponding to positions along the indicated sequence. If outputting in the default "wiggle" format, each line of the file will contain a sequence position (1-based) corresponding to a transposon insertion site and the number of reads aligning to that position separated by a tab. Negative positions represent reads aligned to the left flank (upstream) of the insertion site, positive positions represent reads aligned to the right flank (downstream) of the insertion site.
For sequence reads generated using the transposon insertion sequencing protocol as described by Kazi, Schargel, and Boll. In this process, fastq files should have already been demultiplexed on the instrument into one read file per pool or condition. This script performs the following processing steps:
- Identify and trim transposon sequence from reads
- Align reads to reference sequence(s)
- Assign aligments to positions in reference sequence
- Output read counts
Prerequisites:
bowtie (version 0.12.8 or higher)
Usage:
perl INSeq_read_preprocess_Boll_protocol.pl [options] -r <file of read files> -g <reference_sequence.fasta>
Required arguments:
-r File of demultiplexed read files corresponding to seprate pools, conditions, and/or replicates. Read files must be in fastq format and can be gzipped.
OPTIONAL: Mark input pool(s) with 'i' in a third column. Otherwise will guess input are files that have 'input' (case insensitive) at the beginning of the pool_ID name. This designation be used for insertion site calculations. If the -i option is given, then output read count files will be filtered based on the marked pools (see below).
File should have the following format and be tab delimited with one pool / read file per line:
path/to/read_file.fastq(.gz) pool_ID 'i' or blank (OPTIONAL)
-g Reference genome sequence file, in fasta format. For best results, this file must include sequences for ALL genetic material present in the organism (i.e. chromosomes and plasmids).
Other options:
-t Trim sequence (default: "GGGGACTTATCATCCAACCTGTTA")
-m Number of trim sequence mismatches allowed (default: 1)
-x Maximum number of genome alignment mismatches allowed (default: 1)
-o Output format for count files. Possible options include:
- 'wiggl': Files will be in wiggle format that can be used as input to ESSENTIALS (default)
- 'inseq': Files will be in in the format of *_bowtiemap_processed.txt files produced by Goodman et al's INSeq_analysis software
- 'essen': Files will be in the format of allta_split_*.counts.txt files produced by ESSENTIALS
-s Minimum number of total reads (left flank + right flank) at an insertion site required for output (default: 3)
-i If selected, all output count files will only contain positions and read counts from sites with at least the minumum number of total reads given by -s above in each of the input pool replicates. If no input pools or marked or identified int eh read input file given to -r, this option will be ignored.
-b Path to directory containing bowtie and bowtie-build. For example: /Users/myname/applications/bowtie_folder (default: assumes this directory is in your PATH)
-p Number of parallel threads to run (default: 1)
-P Output stats files prefix (default: 'out'). Output files will be titled:
'.inseq_read_preprocess_stats.txt'
'.inseq_read_preprocess_site_counts.txt'
Output files:
Files will start with a prefix corresponding to the pool_ID given in the input file above.
<prefix>.<reference_sequence>.wiggle: Read alignment counts for each of the individual records that was present in the reference genome sequence file. For example, if the file given to the-roption contained one chromosomal sequence record and one plasmid sequence record there should be two of these files per pool, each named according to the sequence record name and with coordinates corresponding to positions along the indicated sequence. If outputting in the default "wiggle" format, each line of the file will contain a sequence position (1-based) corresponding to a transposon insertion site and the number of reads aligning to that position separated by a tab. Negative positions represent reads aligned to the left flank (upstream) of the insertion site, positive positions represent reads aligned to the right flank (downstream) of the insertion site.
Summary statistics will be output to files and to the screen:
<prefix>.inseq_read_preprocess_stats.txt: Per-file summary statistics
pool: Pool ID
input?: 'Y' = Pool was identifed as an input pool by the user or by pool name
total_reads: Number of reads found in the input read fastq file
reads_w_tn: Number of reads in which the given transposon sequence was found
reads_w_tn_perfect: Number of reads in which the given transposon sequence was found without any mismatch errors
reads_correct_length: Number of reads with the expected legnth after the transposon sequence (23 - 25 bp)
reads_aligned_to_refs: Number of reads with single alignments to any of the reference sequences
reads_aligned_TA: Number of reads with single alignments that aligned at a TA site in a reference sequence
reads_aligned_TA_perfect: Number of reads with single alignments without mismatches that aligned at a TA site in a reference sequence
X_sites: Number of unique TA sites in sequence X with any reads mapped
X_sites_gte_Y: Number of unique TA sites in sequence X with at least Y reads mapped
<prefix>.inseq_read_preprocess_site_counts.txt: Summary of transposon insertion site counts in the given genome sequence(s)
ID: Sequence ID
TA_sites: Total TA motif sites in the sequence
pools: Indicates whether the figures apply to all pools ("all") or just to those identified as input pools ("input")
sites_w_ins: Number of TA sites with at least one read aligning in any of the pools
sites_w_ins_all_pools: Number of TA sites with at least one read aligning in all of the pools
sites_w_ins_gteYrds: Number of TA sites with at least Y reads aligning in any of the pools
sites_w_ins_all_pools_gteYrds: Number of TA sites with at least Y reads aligning in all of the pools