Using BWA Index and i get 5 output files, but BWA-mem asks for just 1

Question

[SaltedPork]

I'm using 1000 genomes pilot data, I've downloaded 3 zipped files. And I then use all three in the same command so it looks like this:

bwa index input_1.fastq input_2.fastq input_3.fastq

This produces 5 files, 2 are binary and 3 are fastq again.

But the usage for BWA-mem asks that the index be specified as 1 file, what's going on here? thanks!

[mirzaabid]

Hi, I was working to make an index file for bacterial genomes by using BWA command

bwa index -a bwtsw bacterial_genomes.fna

This command was working from last 4 days and was functional due to big database, suddenly due to power outage this command was interrupted. I want to continue this command from last point where it was working before power outage. Please help me regarding this issue.

[GenoMax]

No you can't do that. You have to start over. There is no way around this.

Answer 1

[GenoMax]

One does not index fastq files. One would index the genome fasta that you want to search against using the fastq files.

Is that what are you trying to do?

[SaltedPork]

I think so, when you say genome, are you referring to the human genome reference file. I have some 454 data, and a file with LINE-1 transduction data so I'm trying to in simple terms, find all the transduction sites in the 454 data, hence why i thought i would need to index that.

[GenoMax]

A reference can be any fasta sequence. You would index that fasta file and then map your reads against that index. What format is LINE-1 transduction in? Is it a fasta file?

[SaltedPork]

its in fasta. When specifying the reference file in the bwa-mem command, do i use the original fasta, or the newly created 5 files?

[GenoMax]

You will index first.

bwa index -p index_base my_fasta.fa

Then do bwa mem using the "basename" of the index files (which in this case would be index_base, use some other name that you like)

bwa mem [options] index_base file1.fastq

[DNAngel]

Can the reference file for bwa-mem be one file with a list of fasta sequences (i.e. if I wanted to map to 10 diff protein-coding sequences) at once, instead of running 10 different bwa-mem sessions with each different sequence?

[GenoMax]

It can be a multi-fasta file with all 10 sequences in one file. If these sequences are very similar then make sure you keep multi-mapping of reads in mind (so you can recover all hits).

[DNAngel]

Thank you for your response! Some sequences are similar but they differ in length. I have about 20 individual fastq single-end files for my 20 species. Normally if I run bwa mem for 1 CDS, I can obtain an mpileup file - so in a sense I should expect 10 different mpileup files for each line of CDS in my reference file. My test run hasn't ended yet and it seems to be taking a long time, so I think I need to rework my code a bit but when I tested it earlier for a ref.fa file with 2 lines of CDS, I ended up with 1 mpileup file that looked messed up (like the CDS merged together).

I am not sure if bwa mem is reading each line of CDS as a unique reference sequence, but is trying to map to all at once.