Running STAR aligner on paired-end reads as single-end read

Question

[Glory Basumata]

Hi, I am just curious. Can we use STAR aligner to align the paired-end reads as a single-end? What are the consequences of the output of such alignment?

Answer 1

[h.mon]

Yes, STAR can align paired-end reads as single-end, just run STAR two times, one with --readFilesIn R1.fastq, another with --readFilesIn R2.fastq.

What are the consequences of the output of such alignment?

First and foremost, the mapped reads will be considered as single-end. In addition, the proportion of multi-mappers may increase.

[Glory Basumata]

Hi, h.mon, Thank you for your answer. When you say, run STAR two times, do you mean that I take one Fastq file and run it two times? First with Forward option followed by the Reverse alignment option? Thank you.

[h.mon]

First with Forward option followed by the Reverse alignment option?

STAR doens't have Forward or Reverse options.

Assuming you do have paired-end sequencing, your fastq files may be organized in two ways:

• separated in two files, with forward and reverse reads in each file, commonly name_R1.fastq and name_R2.fastq, or some variation of this. This is the more common scenario.

• forward and reverse reads in one file, usually referred to as a interleaved fastq file. This is not so common.

If you have your paired-end reads separated in two files, you have to run STAR twice, once for each file. If you have your paired-end reads in a interleaved fastq file, you can just supply the name of the file and STAR will map it as single-end, as STAR doesn't recognize interleaved files (nor have a parameter to toggle interleaved ON), so the reads will be mapped as single-end.

[Glory Basumata]

Hi h.mon, Thank you for the detailed explanation. I appreciate your help. Sorry for the late response, as I was doubtful when you mentioned that STAR does not have the option to align forward or reverse reads. So I have to take some time and do my research on it. What raised my doubt, initially, was the galaxy project's option for RNA STAR which asks for reads - forward and reverse

https://ibb.co/f18YRU

[h.mon]

Sorry, my explanation was a bit shabby. STAR can accept single-end reads, or paired-end reads in two files. But unlike some programs, where one has parameters for each file separately (e.g. HISAT2, with -1 pair_1.fastq -2 pair_2.fastq), STAR has only one parameter (--readFilesIn), which will behave differently if you pass one or two files to it.

If you pass one file, STAR will consider these as single-end reads: --readFilesIn single_reads.fastq.

If you pass two files, STAR will consider these as paired reads: --readFilesIn pair_1.fastq pair_2.fastq.

Galaxy wraps the command with a web front-end interface, hiding the underlying command-line from users. This means the Galaxy interface can ask for two files separately, but when STAR is run on the server, the argument passed will be --readsFiles pair_1.fastq pair_2.fastq.

[Glory Basumata]

This beautifully crafted answer now clears my doubt altogether in relation to supply of read inputs into STAR or Hisat2. Thank you for your help. Appreciate it.

[xiaoleiusc]

Hi, h.mon, If I run STAR for paired-end reads, if one read in a pair can be aligned and the other one can not be aligned, how does STAR deal with such pair-end reads? Does STAR consider such pair-end reads mapped or unmapped?

Xiao