Exercise #3: PE alignment with BioITeam scripts

Now that you've done everything the hard way, let's see how to do run an alignment pipeline using a BWA alignment script maintained by the BioITeam,  /work2/projects/BioITeam/common/script/align_bwa_illumina.sh. Type in the script name to see its usage.

align_bwa_illumina.sh 2021_06_05
Align Illumina SE or PE data with bwa. Produces a sorted, indexed,
duplicate-marked BAM file and various statistics files. Usage:

align_bwa_illumina.sh <aln_mode> <in_file> <out_pfx> <assembly> [ paired trim_sz trim_sz2 seq_fmt qual_fmt ]

Required arguments:
  aln_mode  Alignment mode, either global (bwa aln) or local (bwa mem).
  in_file   For single-end alignments, path to input sequence file.
            For paired-end alignments using fastq, path to the the R1
            fastq file which must contain the string 'R1' in its name.
            The corresponding 'R2' must have the same path except for 'R1'.
  out_pfx   Desired prefix of output files in the current directory.
  assembly  One of hg38, hg19, hg38, mm10, mm9, sacCer3, sacCer1, ce11, ce10,
            danRer7, hs_mirbase, mm_mirbase, or reference index prefix.
Optional arguments:
  paired    0 = single end alignment (default); 1 = paired end.
  trim_sz   Size to trim reads to. Default 0 (no trimming)
  trim_sz2  Size to trim R2 reads to for paired end alignments.
            Defaults to trim_sz
  seq_fmt   Format of sequence file (fastq, bam or scarf). Default is
            fastq if the input file has a '.fastq' extension; scarf
            if it has a '.sequence.txt' extension.
  qual_type Type of read quality scores (sanger, illumina or solexa).
            Default is sanger for fastq, illumina for scarf.
Environment variables:
  show_only  1 = only show what would be done (default not set)
  aln_args   other bowtie2 options (e.g. '-T 20' for mem, '-l 20' for aln)
  no_markdup 1 = don't mark duplicates (default 0, mark duplicates)
  run_fastqc 1 = run fastqc (default 0, don't run). Note that output
             will be in the directory containing the fastq files.
  keep       1 = keep unsorted BAM (default 0, don't keep)
  bwa_bin    BWA binary to use. Default bwa 0.7.x. Note that bwa 0.6.2
             or earlier should be used for scarf and other short reads.

  align_bwa_illumina.sh local  ABC_L001_R1.fastq.gz my_abc hg38 1
  align_bwa_illumina.sh global ABC_L001_R1.fastq.gz my_abc hg38 1 50
  align_bwa_illumina.sh global sequence.txt old sacCer3 0 '' '' scarf solexa

There are lots of bells and whistles in the arguments, but the most important are the first few:

  1. aln_mode – whether to perform a global or local alignment (the 1st argument must be one of those words)
  2. in_file – full or relative path to the FASTQ file (just the R1 fastq if paired end). Can be compressed (.gz)
  3. out_pfxprefix for all the output files produced by the script. Should relate back to what the data is.
  4. assembly – genome assembly to use.
  5. paired flag 0 means single end (the default); 1 means paired end
  6. trim_sz – if you want the FASTQ hard trimmed down to a specific length before alignment, supply that number here

We're going to run this script and a similar Bowtie2 alignment script, on the yeast data using the TACC batch system. In a new directory, copy over the commands and submit the batch job. We ask for 2 hours (-t 02:00:00) with 4 tasks/node (-w 4); since we have 4 commands, this will run on 1 compute node.

# Make sure you're not in an idev session by looking at the hostname
# If the hostname looks like "c455-004.stampede2.tacc.utexas.edu", exit the idev session

# Copy over the Yeast data if needed
mkdir -p $SCRATCH/core_ngs/alignment/fastq
cp $CORENGS/alignment/Sample_Yeast*.gz $SCRATCH/core_ngs/alignment/fastq/

# Make a new alignment directory for running these scripts
mkdir -p $SCRATCH/core_ngs/alignment/bwa_script
cd $SCRATCH/core_ngs/alignment/bwa_script
ln -s -f ../fastq

# Copy the alignment commands file and submit the batch job
cp $CORENGS/tacc/aln_script.cmds .
launcher_creator.py -j aln_script.cmds -n aln_script -t 02:00:00 -w 4 -a UT-2015-05-18 -q normal
sbatch --reservation=BIO_DATA_week_1 aln_script.slurm 
showq -u

While we're waiting for the job to complete, lets look at the aln_script.cmds file.

/work2/projects/BioITeam/common/script/align_bwa_illumina.sh     global ./fastq/Sample_Yeast_L005_R1.cat.fastq.gz bwa_global sacCer3 1 50
/work2/projects/BioITeam/common/script/align_bwa_illumina.sh     local  ./fastq/Sample_Yeast_L005_R1.cat.fastq.gz bwa_local  sacCer3 1
/work2/projects/BioITeam/common/script/align_bowtie2_illumina.sh global ./fastq/Sample_Yeast_L005_R1.cat.fastq.gz bt2_global sacCer3 1 50
/work2/projects/BioITeam/common/script/align_bowtie2_illumina.sh local  ./fastq/Sample_Yeast_L005_R1.cat.fastq.gz bt2_local  sacCer3 1


Output files

This alignment pipeline script performs the following steps:

There are a number of output files, with the most important being those desribed below.

  1. <prefix>.align.log – Log file of the entire alignment process.
  2. <prefix>.sort.dup.bam – Sorted, duplicate-marked alignment file.
  3. <prefix>.sort.dup.bam.bai – Index for the sorted, duplicate-marked alignment file
  4. <prefix>.flagstat.txt samtools flagstat output
  5. <prefix>.idxstats.txt samtools idxstats output
  6. <prefix>.samstats.txt – Summary alignment statistics from Anna's stats script
  7. <prefix>.iszinfo.txt – Insert size statistics (for paired-end alignments) from Anna's stats script

Verifying alignment success

The alignment log will have a  "I ran successfully" message at the end if all went well, and if there was an error, the important information should also be at the end of the log file. So you can use tail to check the status of an alignment. For example:

tail bwa_global.align.log

This will show something like:

..Done alignmentUtils.pl bamstats - 2020-06-14 23:19:38
.. samstats file 'bwa_global.samstats.txt' exists and is not empty - 2020-06-14 23:19:38
## Cleaning up files (keep 0) - 2020-06-14 23:19:38
ckRes 0 cleanup
## All bwa alignment tasks completed successfully! - 2020-06-14 23:19:38

Notice that success message:  "All bwa alignment tasks completed successfully!". It should only appear once in any successful alignment log.

When multiple alignment commands are run in parallel it is important to check them all, and you can use grep looking for part of the unique success message to do this. For example:

grep 'completed successfully!' *align.log | wc -l

If this command returns 4 (the number of alignment tasks we performed), all went well, and we're done.

But what if something went wrong? How can we tell which alignment task was not successful? You could tail the log files one by one to see which one(s) don't have the message, but you can also use a special grep option to do this work.

grep -L 'completed successfully' *.align.log

The -L option tells grep to only print the filenames that don't contain the pattern. Perfect! To see happens in the case of failure, try it on a file that doesn't contain that message:

grep -L 'completed successfully' aln_script.cmds

Checking alignment statistics

The <prefix>.samstats.txt statistics files produced by the alignment pipeline has a lot of good information in one place. If you look at bwa_global.samstats.txt you'll see something like this:

             Aligner:       bwa
     Total sequences:   1184360
        Total mapped:    539079 (45.5 %)
      Total unmapped:    645281 (54.5 %)
             Primary:    539079 (100.0 %)
          Duplicates:    249655 (46.3 %)
          Fwd strand:    267978 (49.7 %)
          Rev strand:    271101 (50.3 %)
          Unique hit:    503629 (93.4 %)
           Multi hit:     35450 (6.6 %)
           Soft clip:
           All match:    531746 (98.6 %)
              Indels:      7333 (1.4 %)
       Total PE seqs:   1184360
      PE seqs mapped:    539079 (45.5 %)
        Num PE pairs:    592180
   F5 1st end mapped:    372121 (62.8 %)
   F3 2nd end mapped:    166958 (28.2 %)
     PE pairs mapped:     80975 (13.7 %)
     PE proper pairs:     16817 (2.8 %)

Since this was a paired end alignment there is paired-end specific information reported.

You can also view statistics on insert sizes for properly paired reads in the bwa_global.iszinfo.txt file. This tells you the average (mean) insert size, standard deviation, mode (most common value), and fivenum values (minimum, 1st quartile, median, 3rd quartile, maximum).

  Insert size stats for: bwa_global
        Number of pairs: 16807 (proper)
 Number of insert sizes: 406
        Mean [-/+ 1 SD]: 296 [176 416]  (sd 120)
         Mode [Fivenum]: 228  [51 224 232 241 500]

A quick way to check alignment stats if you have run multiple alignments is again to use grep. For example:

grep 'Total mapped' *samstats.txt

will produce output like this:

bt2_global.samstats.txt:        Total mapped:    602893 (50.9 %)
bt2_local.samstats.txt:        Total mapped:    788069 (66.5 %)
bwa_global.samstats.txt:        Total mapped:    539079 (45.5 %)
bwa_local.samstats.txt:        Total mapped:   1008000 (76.5 %

Exercise: How would you list the median insert size for all the alignments?

That information is in the *.iszinfo.txt files, on the line labeled Mode.

The median value is th 3rd value in the 5 fivnum values; it is the 7th whitespace-separated field on the Mode line.

Use grep to isolate the Mode line, and awk to isolate the median value field:

grep 'Mode' *.iszinfo.txt | awk '{print $1,"Median insert size:",$7}'

TACC batch system considerations

The great thing about pipeline scripts like this is that you can perform alignments on many datasets in parallel at TACC, and they are written to take advantage of having multiple cores on TACC nodes where possible.

On the stampede2, with its 68 physical cores per node, they are designed to run best with no more than 4 tasks per node.

These alignment scripts should always be run with a wayness of 4 (-w 4) in the stampede2 batch system, meaning at most 4 commands per node.

Exercise #4: Bowtie2 global alignment - Vibrio cholerae RNA-seq

While we have focused on aligning eukaryotic data, the same tools can be used with prokaryotic data. The major differences are less about the underlying data and much more about the external/public databases that store and distribute reference data. If we want to study a prokaryote, the reference data is usually downloaded from a resource like GenBank.  

In this exercise, we will use some RNA-seq data from Vibrio cholerae, published on GEO here, and align it to a reference genome.

Overview of Vibrio cholerae alignment workflow with Bowtie2

Alignment of this prokaryotic data follows the workflow below. Here we will concentrate on steps 1 and 2.

  1. Prepare the vibCho reference index for bowtie2 from GenBank records
  2. Align reads using bowtie2, producing a SAM file
  3. Convert the SAM file to a BAM file (samtools view) 
  4. Sort the BAM file by genomic location (samtools sort)
  5. Index the BAM file (samtools index)
  6. Gather simple alignment statistics (samtools flagstat and samtools idxstat)

Obtaining the GenBank records

First prepare a directory to work in, and change to it:

mkdir -p $SCRATCH/core_ngs/references/vibCho
cd $SCRATCH/core_ngs/references/vibCho

V. cholerae has two chromosomes. We download each separately.

  1. Navigate to http://www.ncbi.nlm.nih.gov/nuccore/NC_012582
  2. Back on the main http://www.ncbi.nlm.nih.gov/nuccore/NC_012582 page
  3. Repeat steps 1 and 2 for the 2nd chromosome
  4. Transfer the files from your local computer to TACC

mkdir -p $SCRATCH/core_ngs/references/vibCho
cd $SCRATCH/core_ngs/references/vibCho
cp $CORENGS/ncbi/vibCho/NC_* .

Once you have the 4 files locally in your $SCRATCH/core_ngs/references/vibCho directory, combine them using cat:

cd $SCRATCH/core_ngs/references/vibCho
cat NC_01258[23].fa   > vibCho.O395.fa
cat NC_01258[23].gff3 > vibCho.O395.gff3

Now we have a reference sequence file that we can use with the bowtie2 reference builder, and ultimately align sequence data against.

Introducing bowtie2

First make sure you're in an idev session:

idev -m 180 -p normal -A UT-2015-05-18 -N 1 -n 68 

Go ahead and load the bowtie2 module so we can examine some help pages and options.

module biocontainers
module load bowtie

 Now that it's loaded, check out the options. There are a lot of them! In fact for the full range of options and their meaning, Google "Bowtie2 manual" and bring up that page (http://bowtie-bio.sourceforge.net/bowtie2/manual.shtml). The Table of Contents is several pages long! Ouch!

This is the key to using bowtie2 - it allows you to control almost everything about its behavior, which make it the go-to aligner for specialized alignment tasks (e.g. aligning miRNA or other small reads). But it also makes it is much more challenging to use than bwa – and it's easier to screw things up too!

Building the bowtie2 vibCho index

Before the alignment, of course, we've got to build a bowtie2- specific index using bowtie2-build. Go ahead and check out its options. Unlike for the aligner itself, we only need to worry about a few things here:

Here, to build the reference index for alignment, we only need the FASTA file. (This is not always true - extensively spliced transcriptomes requires splice junction annotations to align RNA-seq data properly.)

First create a directory specifically for the bowtie2 index, then build the index using bowtie-build.

mkdir -p $SCRATCH/core_ngs/references/bt2/vibCho
cd $SCRATCH/core_ngs/references/bt2/vibCho

# Symlink to the fasta file you created
ln -sf $SCRATCH/core_ngs/references/vibCho.O395.fa
# or, to catch up:
ln -sf $CORENGS/references/vibCho.O395.fa

bowtie2-build vibCho.O395.fa vibCho.O395

This should also go pretty fast. You can see the resulting files using ls like before.

Performing the bowtie2 alignment

We'll set up a new directory to perform the V. cholerae data alignment. But first make sure you have the FASTQ file to align and the vibCho bowtie2 index:

# Get the FASTQ to align
mkdir -p $SCRATCH/core_ngs/alignment/fastq
cp $CORENGS/alignment/*fastq.gz $SCRATCH/core_ngs/alignment/fastq/

# Set up the bowtie2 index
mkdir -p $SCRATCH/core_ngs/references/bt2/vibCho
cp $CORENGS/idx/bt2/vibCho/*.*  $SCRATCH/core_ngs/references/bt2/vibCho

Make sure you're in an idev session with the bowtie2 BioContainers module loaded:

idev -p normal -m 120 -A UT-2015-05-18 -N 1 -n 68
module load biocontainers
module load bowtie

Now set up a directlry to do this alignment, with symbolic links to the bowtie2 index directory and the directory containing the FASTQ to align:

mkdir -p $SCRATCH/core_ngs/alignment/vibCho
cd $SCRATCH/core_ngs/alignment/vibCho
ln -sf ../../references/bt2/vibCho
ln -sf ../../alignment/fastq fq

We'll be aligning the V. cholerae reads now in ./fq/cholera_rnaseq.fastq.gz (how many sequences does it contain?)

Note that here the data is from standard mRNA sequencing, meaning that the DNA fragments are typically longer than the reads. There is likely to be very little contamination that would require using a local rather than global alignment, or many other pre-processing steps (e.g. adapter trimming). Thus, we will run bowtie2 with default parameters, omitting options other than the input, output, and reference index. This performs a global alignment.

As you can tell from looking at the bowtie2 help message, the general syntax looks like this:

bowtie2 [options]* -x <bt2-idx> {-1 <m1> -2 <m2> | -U <r>} [-S <sam>]

So execute this bowtie2 global, single-end alignment command:

cd $SCRATCH/core_ngs/alignment/vibCho
bowtie2 -x vibCho/vibCho.O395 -U fq/cholera_rnaseq.fastq.gz -S cholera_rnaseq.sam 2>&1 | tee aln_global.log


Since the FASTQ file is not large, this should not take too long, and you will see progress output like this:

89006 reads; of these:
  89006 (100.00%) were unpaired; of these:
    20675 (23.23%) aligned 0 times
    38226 (42.95%) aligned exactly 1 time
    30105 (33.82%) aligned >1 times
76.77% overall alignment rate

When the job is complete you should have a cholera_rnaseq.sam file that you can examine using whatever commands you like.  Remember, to further process it downstream, you should create a sorted, indexed BAM file from this SAM output.

Exercise: Repeat the alignment performing a local alignment, and write the output in BAM format. How do the alignment statistics compare?


module load samtools
cd $SCRATCH/core_ngs/alignment/vibCho
bowtie2 --local -x vibCho/vibCho.O395 -U fq/cholera_rnaseq.fastq.gz 2>aln_local.log | \
  samtools view -b > cholera_rnaseq.local.bam

Reports these alignment statistics:

89006 reads; of these:
  89006 (100.00%) were unpaired; of these:
    27061 (30.40%) aligned 0 times
    33828 (38.01%) aligned exactly 1 time
    28117 (31.59%) aligned >1 times
69.60% overall alignment rate

Interestingly, the local alignment rate here is lower than we saw with the gloabl alignment.

Exercise #5: BWA-MEM - Human mRNA-seq

After bowtie2 came out with a local alignment option, it wasn't long before bwa developed its own local alignment algorithm called BWA-MEM (for Maximal Exact Matches), implemented by the bwa mem command. bwa mem has the following advantages:

This exercise will align a human total RNA-seq dataset composed (by design) almost exclusively of reads that cross splice junctions.

A word about real splice-aware aligners

Using BWA mem for RNA-seq alignment is sort of a "poor man's" RNA-seq alignment method. Real splice-aware aligners like tophat2, hisat2 or STAR have more complex algorithms (as shown below) – and take a lot more time!

In the transcriptome-aware alignment above, reads that span splice junctions are reported in the SAM file with genomic coordinates that start in the first exon and end in the second exon (the CIGAR string uses the N operator, e.g. 30M1000N60M).

BWA MEM does not know about the exon structore of the genome. But it can align different sub-sections of a read to two different locations, producing two alignment records from one input read (one of the two will be marked as secondary (0x100 flag).

BWA MEM splits junction-spanning reads into two alignment records

Setup for BWA mem

First set up our working directory for this alignment. Since it takes a long time to build a bwa index for a large genome (here human hg38/GRCh38), we'll use one that the BioITeam maintains in its /work2/projects/BioITeam/ref_genome area.

# Make sure you're in an idev session
idev -m 120 -p normal -A UT-2015-05-18 -N 1 -n 68

# Load the modules we'll need
module load biocontainers
module load bwa
module load samtools

# Copy over the FASTQ data if needed
mkdir -p $SCRATCH/core_ngs/alignment/fastq
cp $CORENGS/alignment/*.gz $SCRATCH/core_ngs/alignment/fastq/

# Make a new alignment directory for running these scripts
mkdir -p $SCRATCH/core_ngs/alignment/bwamem
cd $SCRATCH/core_ngs/alignment/bwamem
ln -sf ../fastq
ln -sf /work2/projects/BioITeam/ref_genome/bwa/bwtsw/hg38

Now take a look at bwa mem usage (type bwa mem with no arguments). The most important parameters are the following:

-kControls the minimum seed length (default = 19)
-wControls the "gap bandwidth", or the length of a maximum gap. This is particularly relevant for MEM, since it can determine whether a read is split into two separate alignments or is reported as one long alignment with a long gap in the middle (default = 100)
-MFor split reads, mark the shorter read as secondary
-rControls how long an alignment must be relative to its seed before it is re-seeded to try to find a best-fit local match (default = 1.5, e.g. the value of -k multiplied by 1.5)
-cControls how many matches a MEM must have in the genome before it is discarded (default = 10000)
-tControls the number of threads to use

RNA-seq alignment with bwa mem

Based on its help info, this is the structure of the bwa mem command we will use:

bwa mem -M <ref.fa> <reads.fq> > outfile.sam

After performing the setup above, execute the following command in your idev session:

cd $SCRATCH/core_ngs/alignment/bwamem
bwa mem -M hg38/hg38.fa fastq/human_rnaseq.fastq.gz 2>hs_rna.bwamem.log |
  samtools view -b | \
  samtools sort -O BAM -T human_rnaseq.tmp > human_rnaseq.sort.bam

This multi-pipe command performs three steps:

Because the progress output is being redirected to a log file, we won't see anything until the command completes. Then you should have a human_rnaseq.sort.bam file and an hs_rna.bwamem.log logfile.

Exercise: Compare the number of original FASTQ reads to the number of alignment records.

Use the zcat | wc -l  | awk idiom to count FASTQ reads.

Use samtools flagstat to report alignment statistics.

Count the FASTQ file reads:

cd $SCRATCH/core_ngs/alignment/bwamem
zcat ./fastq/human_rnaseq.fastq.gz | wc -l | awk '{print $1/4}'

The file has 100,000 reads.

Generate alignment statistics from the sorted BAM file:

cd $SCRATCH/core_ngs/alignment/bwamem
samtools flagstat human_rnaseq.sort.bam | tee hs_rnaseq.flagstat.txt

Output will look like this:

133570 + 0 in total (QC-passed reads + QC-failed reads)
33570 + 0 secondary
0 + 0 supplementary
0 + 0 duplicates
133450 + 0 mapped (99.91% : N/A)
0 + 0 paired in sequencing
0 + 0 read1
0 + 0 read2
0 + 0 properly paired (N/A : N/A)
0 + 0 with itself and mate mapped
0 + 0 singletons (N/A : N/A)
0 + 0 with mate mapped to a different chr
0 + 0 with mate mapped to a different chr (mapQ>=5)

There were 133,570 alignment records reported for the 100,000 input reads.

Because bwa mem can split reads and report two alignment records for the same read, there are 33,570 secondary reads reported here.

Be aware that some downstream tools (for example the Picard suite, often used before SNP calling) do not like it when a read name appears more than once in the SAM file. Such reads can be filtered, but only if they can be identified as secondary by specifying the bwa mem -M option as we did above. This option leaves the longest alignmen normally but marks additional alignments for the read as secondary (the 0x100 BAM flag). This designation also allows you to easily filter the secondary reads with samtools view -F 0x104 if desired.