Tweet
Hello,
If you are dabbling into transcriptome assemblers like Oases or Trinity, I wrote few commentaries that you may find useful.
I have not got time to finish writing on algorithms of Trinity/Oases in detail, but that post is coming soon. These posts on de Bruijn graph series (http://www.homolog.us/blogs/category/de-bruijn/) cover many of the basic concepts, but I think it will help everyone go through the algorithms of specific transcriptome assemblers.
One point try to get across is that you should think more in terms of how many independent k-mers you have rather than how many reads you have. The k-mer content of your library is what determines RAM size. If you create a simulated library of 100 M reads from one gene, the RAM size requirement will be minimal for a well-crafted de Bruijn assembly algorithm.
If you are dabbling into transcriptome assemblers like Oases or Trinity, I wrote few commentaries that you may find useful.
I have not got time to finish writing on algorithms of Trinity/Oases in detail, but that post is coming soon. These posts on de Bruijn graph series (http://www.homolog.us/blogs/category/de-bruijn/) cover many of the basic concepts, but I think it will help everyone go through the algorithms of specific transcriptome assemblers.
One point try to get across is that you should think more in terms of how many independent k-mers you have rather than how many reads you have. The k-mer content of your library is what determines RAM size. If you create a simulated library of 100 M reads from one gene, the RAM size requirement will be minimal for a well-crafted de Bruijn assembly algorithm.