Read Mapping
During the next generation sequencing portion of our workflow, it's as if a box of puzzle pieces were dumped out onto a table. To get a full picture of an individual's genome, we need to put that puzzle back together through a process called read mapping. As with any puzzle, you need the picture on the front of the box that describes what the puzzle should look like. In genomics, this proverbial "picture on the front of the box" is called the reference genome. You can learn more about the reference genome here.
Before starting your analysis, you'll need to decide on which reference genome you will use as the baseline for your project. Once a reference genome is decided, the next step is to choose a mapping software (frequently referred to as a mapper or aligner) to align the reads to that reference genome. The mapping software generates a special data structure from a reference genome and then goes through each read pair in the FASTQ files to see if and where they fit in the reference genome.
At the time of writing, the most popular aligner for each type of data is BWA for DNA sequencing and STAR for RNA sequencing. Each program provides many parameters that can be tweaked depending on the characteristics of the data to achieve the "best" alignment. Once the alignment program has completed, it creates a SAM/BAM file containing all of the read pairs, where the aligner believes they live in the reference genome (if anywhere), and many other fields that the aligner computes. This file is widely considered the basis of most modern computational genomic analyses. The specifics of each file type are discussed in the next chapter.
As an implementation note, the entire analysis is based off of the reference genome used. Generally speaking, analyses that use two different reference genomes are not compatible because the positions of genes and other important biomarkers are updated with each genome build. Once an end-to-end analysis pipeline is decided on, it is a significant task to update analysis pipelines to a new reference genome. As such, at the time of writing, many labs are still using the GRCh37 version of the human genome, which was released in 2009.