RNA-seq Bioinformatics

Introduction to bioinformatics for RNA sequence analysis

Alignment

HISAT2 alignment

Perform alignments with HISAT2 to the genome and transcriptome.

First, begin by making the appropriate output directory for our alignment results.

    echo $RNA_ALIGN_DIR
    mkdir -p $RNA_ALIGN_DIR
    cd $RNA_ALIGN_DIR

HISAT2 uses a graph-based alignment and has succeeded HISAT and TOPHAT2. The output of this step will be a SAM/BAM file for each data set.

Refer to HISAT2 manual for a more detailed explanation:

HISAT2 basic usage:

    #hisat2 [options]* -x <ht2-idx> {-1 <m1> -2 <m2> | -U <r> | --sra-acc <SRA accession number>} [-S <sam>]

Extra options specified below:

hisat2 -p 8 --rg-id=UHR_Rep1 --rg SM:UHR --rg LB:UHR_Rep1_ERCC-Mix1 --rg PL:ILLUMINA --rg PU:CXX1234-ACTGAC.1 -x $RNA_REF_INDEX --dta --rna-strandness RF -1 $RNA_DATA_DIR/UHR_Rep1_ERCC-Mix1_Build37-ErccTranscripts-chr22.read1.fastq.gz -2 $RNA_DATA_DIR/UHR_Rep1_ERCC-Mix1_Build37-ErccTranscripts-chr22.read2.fastq.gz -S ./UHR_Rep1.sam
hisat2 -p 8 --rg-id=UHR_Rep2 --rg SM:UHR --rg LB:UHR_Rep2_ERCC-Mix1 --rg PL:ILLUMINA --rg PU:CXX1234-TGACAC.1 -x $RNA_REF_INDEX --dta --rna-strandness RF -1 $RNA_DATA_DIR/UHR_Rep2_ERCC-Mix1_Build37-ErccTranscripts-chr22.read1.fastq.gz -2 $RNA_DATA_DIR/UHR_Rep2_ERCC-Mix1_Build37-ErccTranscripts-chr22.read2.fastq.gz -S ./UHR_Rep2.sam
hisat2 -p 8 --rg-id=UHR_Rep3 --rg SM:UHR --rg LB:UHR_Rep3_ERCC-Mix1 --rg PL:ILLUMINA --rg PU:CXX1234-CTGACA.1 -x $RNA_REF_INDEX --dta --rna-strandness RF -1 $RNA_DATA_DIR/UHR_Rep3_ERCC-Mix1_Build37-ErccTranscripts-chr22.read1.fastq.gz -2 $RNA_DATA_DIR/UHR_Rep3_ERCC-Mix1_Build37-ErccTranscripts-chr22.read2.fastq.gz -S ./UHR_Rep3.sam

hisat2 -p 8 --rg-id=HBR_Rep1 --rg SM:HBR --rg LB:HBR_Rep1_ERCC-Mix2 --rg PL:ILLUMINA --rg PU:CXX1234-TGACAC.1 -x $RNA_REF_INDEX --dta --rna-strandness RF -1 $RNA_DATA_DIR/HBR_Rep1_ERCC-Mix2_Build37-ErccTranscripts-chr22.read1.fastq.gz -2 $RNA_DATA_DIR/HBR_Rep1_ERCC-Mix2_Build37-ErccTranscripts-chr22.read2.fastq.gz -S ./HBR_Rep1.sam
hisat2 -p 8 --rg-id=HBR_Rep2 --rg SM:HBR --rg LB:HBR_Rep2_ERCC-Mix2 --rg PL:ILLUMINA --rg PU:CXX1234-GACACT.1 -x $RNA_REF_INDEX --dta --rna-strandness RF -1 $RNA_DATA_DIR/HBR_Rep2_ERCC-Mix2_Build37-ErccTranscripts-chr22.read1.fastq.gz -2 $RNA_DATA_DIR/HBR_Rep2_ERCC-Mix2_Build37-ErccTranscripts-chr22.read2.fastq.gz -S ./HBR_Rep2.sam
hisat2 -p 8 --rg-id=HBR_Rep3 --rg SM:HBR --rg LB:HBR_Rep3_ERCC-Mix2 --rg PL:ILLUMINA --rg PU:CXX1234-ACACTG.1 -x $RNA_REF_INDEX --dta --rna-strandness RF -1 $RNA_DATA_DIR/HBR_Rep3_ERCC-Mix2_Build37-ErccTranscripts-chr22.read1.fastq.gz -2 $RNA_DATA_DIR/HBR_Rep3_ERCC-Mix2_Build37-ErccTranscripts-chr22.read2.fastq.gz -S ./HBR_Rep3.sam

Note: in the above alignments, we are treating each library as an independent data set. If you had multiple lanes of data for a single library, you could align them all together in one HISAT2 command. Similarly you might combine technical replicates into a single alignment run (perhaps after examining them and removing outliers…). To combine multiple lanes, you would provide all the read1 files as a comma separated list for the ‘-1’ input argument, and then all read2 files as a comma separated list for the ‘-2’ input argument, (where both lists have the same order) : You can also use samtools merge to combine bam files after alignment. This is the approach we will take.

HISAT2 Alignment Summary HISAT2 generates a summary of the alignments printed to the terminal. Notice the number of total reads, reads aligned and various metrics regarding how the reads aligned to the reference.

SAM to BAM Conversion Convert HISAT2 sam files to bam files and sort by aligned position

    samtools sort -@ 8 -o UHR_Rep1.bam UHR_Rep1.sam
    samtools sort -@ 8 -o UHR_Rep2.bam UHR_Rep2.sam
    samtools sort -@ 8 -o UHR_Rep3.bam UHR_Rep3.sam
    samtools sort -@ 8 -o HBR_Rep1.bam HBR_Rep1.sam
    samtools sort -@ 8 -o HBR_Rep2.bam HBR_Rep2.sam
    samtools sort -@ 8 -o HBR_Rep3.bam HBR_Rep3.sam

Merge HISAT2 BAM files

Make a single BAM file combining all UHR data and another for all HBR data. Note: This could be done in several ways such as ‘samtools merge’, ‘bamtools merge’, or using picard-tools (see below). We chose the third method because it did the best job at merging the bam header information. NOTE: sambamba also retains header info.

    cd $RNA_HOME/alignments/hisat2
    java -Xmx2g -jar $RNA_HOME/tools/picard.jar MergeSamFiles OUTPUT=UHR.bam INPUT=UHR_Rep1.bam INPUT=UHR_Rep2.bam INPUT=UHR_Rep3.bam
    java -Xmx2g -jar $RNA_HOME/tools/picard.jar MergeSamFiles OUTPUT=HBR.bam INPUT=HBR_Rep1.bam INPUT=HBR_Rep2.bam INPUT=HBR_Rep3.bam

Count the alignment (BAM) files to make sure all were created successfully (you should have 8 total)

    ls -l *.bam | wc -l
    ls -l *.bam

PRACTICAL EXERCISE 6

Assignment: Perform some alignments on additional read data sets. Align the reads using the skills you learned above. Try using the HISAT2 aligner. Also practice converting SAM to BAM files, and merging BAM files.

Hint: Do this analysis on the additional data and in the separate working directory called ‘practice’ that you created in Practical Exercise 3. Questions

What is the difference between a .sam and .bam file? If you sorted the resulting BAM file as we did above, is the result sorted by read name? Or position? Which columns of the BAM file can be viewed to determine the style of sorting? What command can you use to view only the BAM header?

Solution: When you are ready you can check your approach against the Solutions.