TL;DR: It’s fine for a very rough estimate.

There are at least two factors that complicate using the filesize of a Fasta file as a proxy for the genome size. First, there are extra characters in the file that do not represent nucleotides in the genome. A small number of these are found in the Fasta record headers, but most are in the form of invisible newline characters. Using the number of bytes in the file as an estimate of genome size will result in an inflated estimate unless the newlines (and other extra characters) are first removed. Second, when working with large Fasta files, it’s common to discuss things in terms of megabytes or gigabytes, not millions or billions of bytes. This is another complication, since 1 megabyte = 1024 kilobytes = 1024 * 1024 bytes, whereas 1 megabase = 1000 kilobases = 1000 * 1000 base pairs.

All of this of course assumes that the Fasta file contains an accurate assembly of the genome. Of course no assembly is perfectly accurate, but bigger problems with the assembly will yield bigger differences between the actual genome size and the genome size estimated by the Fasta file.

That being said, if you’re interested in just a rough estimate, looking at the filesize will give you a ballpark idea of the filesize. Consider the following.

[standage@gnomic Polistes_dominulus]$ perl seqlendist.pl < allpathslg05-final.assembly.fasta | perl -e '$total = 0; while(<>){chomp();($id, $length) = split(/,/); $total += $length;}; printf("length: %d\n", $total)'
length: 202344795
[standage@gnomic Polistes_dominulus]$ ls -l allpathslg05-final.assembly.fasta
-rw-------. 1 standage users 204905495 Apr 21 20:30 allpathslg05-final.assembly.fasta
[standage@gnomic Polistes_dominulus]$ ls -lh allpathslg05-final.assembly.fasta
-rw-------. 1 standage users 196M Apr 21 20:30 allpathslg05-final.assembly.fasta

The first command gives the actual size of the genome assembly: 202,344,795 base pairs. The second command gives the size of the Fasta file containing the assembly: 204,905,495 bytes. The third command gives the human-readable size of the Fasta file: 196 megabytes. In each case, as long as you do some generous rounding, you’ll end up with 200 megabases as your estimate.

Processing DNA or protein sequence files is one of the most common tasks in bioinformatics. Occasionally, the need comes up to split a Fasta file into smaller pieces. This is a task for which I have written many a 5-line Perl script that I used only once. But after today, I don’t think I’ll be writing any more.

I’m a big fan of the GenomeTools toolkit and development library, and today I discovered their splitfasta tool. If you want to split your sequence data into a given number of files, run it with the -numfiles flag. If you want to split your data into files of a certain size, run it with the -targetsize flag. splitfasta usually gives a pretty even distribution of sequence, although occasionally the last file is significantly smaller than the rest–especially when using the -targetsize flag. I saw similar behavior in the scripts that I wrote previously though, so I’m not too concerned.

Just like everything else in the GenomeTools toolkit, splitfasta is implemented in C, so it’s fast. Enjoy!

standage@ubuntu:~/$ ls -lhp
total 98M
-rw-r--r-- 1 standage standage 98M 2012-01-20 08:50 cotton.fasta
standage@ubuntu:~/$ gt splitfasta -numfiles 13 cotton.fasta 
standage@ubuntu:~/$ ls -lhp
total 195M
-rw-r--r-- 1 standage standage  98M 2012-01-20 08:50 cotton.fasta
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.1
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.10
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.11
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.12
-rw-rw-r-- 1 standage standage 7.4M 2012-01-20 10:10 cotton.fasta.13
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.2
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.3
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.4
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.5
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.6
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.7
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.8
-rw-rw-r-- 1 standage standage 7.5M 2012-01-20 10:10 cotton.fasta.9
standage@ubuntu:~/$ rm cotton.fasta.*
standage@ubuntu:~/$ gt splitfasta -targetsize 15 cotton.fasta 
standage@ubuntu:~/$ ls -lhp
total 195M
-rw-r--r-- 1 standage standage  98M 2012-01-20 08:50 cotton.fasta
-rw-rw-r-- 1 standage standage  16M 2012-01-20 10:11 cotton.fasta.1
-rw-rw-r-- 1 standage standage  16M 2012-01-20 10:11 cotton.fasta.2
-rw-rw-r-- 1 standage standage  16M 2012-01-20 10:11 cotton.fasta.3
-rw-rw-r-- 1 standage standage  16M 2012-01-20 10:11 cotton.fasta.4
-rw-rw-r-- 1 standage standage  16M 2012-01-20 10:11 cotton.fasta.5
-rw-rw-r-- 1 standage standage  16M 2012-01-20 10:11 cotton.fasta.6
-rw-rw-r-- 1 standage standage 7.1M 2012-01-20 10:11 cotton.fasta.7