In UPSCb/RnaSeqTutorial: RNA-Seq Tutorial

Visualizing BAM files

Introduction

The integrative genomics viewer (IGV) is a tool which enables its users to see a visual representation of a wide variety of data commonly used in bioinformatics pipelines [@Robinson2011]. In this short tutorial we are going to load two BAM (binary alignment map) files into IGV and inspect a region of interest.

Practical

First, download the relevant files from the server, by opening a browser and navigating to

watson.plantphys.umu.se:82`$PORT`

where $PORT is your assigned port.

You will need:

1. the alignment file (BAM) and index (BAI)

2. the genome FASTA file

You can now download IGV from http://www.broadinstitute.org/igv/projects/current/igv_mm.jnlp. After opening the application, we can start by loading a reference genome via the Genomes > Load Genome from File menu items. In the “File Name” field, enter the Downloads directory of your computer and press return. In this directory, you see a FASTA (.fa) and a FASTA-index (.fai) file of the reference genome for our organism P. trichocarpa. Select the FASTA file and click “Open”. After the reference is loaded, we can proceed by loading the alignments files. Select File > Load from File from the menu and navigate to the aligned files by entering just that in the “File Name” field (confirm again with return). Select the .bam (not the .bai, these are the BAM-index files which we created earlier) file. Confirm with “Open”. In order for us to see the actual alignments, we need to zoom in to a region. Our region of interest is a gene that lies between nucleotides 6553903 and 6561936 on chromosome 19, therefore we enter “Chr19:6553903-6561936” in the gene locus field (at the top of the menu with a “Go” button attached to it).

Discuss what you see. Are there obvious differences between the samples? What do the colours within the reads represent?

Reads are represented as grey boxes with a sharp tip indicating directionality. Small lines indicate that a read spans this area (e.g.: introns). Coloured lines in the reads represent a deviation from the reference, commonly SNPs. Hovering over the coloured line brings up a small dialog with additional information. More info can be found in the View > Color Legends menu.

A glimpse of the future

Not too far distant, just a few pages away. Although this is not per se a data preparation step, we advise at this stage to conduct a number of analyses to assess the biological soundness of the data, such as examining how well biological replicates correlate, how the samples cluster in a principal component analysis (PCA) and whether the first dimensions of the PCA can likely be explained by the biological factors under consideration. To achieve this, it is important to have first normalized the data. When a sufficient number of replicates per condition are available (at least three) we recommend that the data be normalized using a Variance Stabilizing Transformation (VST) such as that implemented in the R/Bioconductor DESeq2 package [@Love:2014p6358], otherwise the data should be normalized using other approaches such as those implemented in the edgeR [@pmid19910308] or DESeq2 packages, approaches assuming a negative binomial distribution of the data.

This concludes the preprocessing part of this tutorial. You now have data ready for the actual analysis.

UPSCb/RnaSeqTutorial documentation built on Nov. 24, 2020, 12:40 a.m.