I suppose you should decide what is the purpose of the alignment you are going to carry out and what datasets are used as a query and a subject. It will determine the alignment tools to be used.

Tools like LASTZ or LAST perform whole-genome alignments of large assembled genome fragments (chromosomes, scaffolds, contigs). They are optimized in the way to create so-called chained alignments, that is, alignments which consist of gap-free alignment blocks separated by large gaps. Such alignments are typically used in studying genome rearrangements, synteny blocks and homologous genome regions.

Alignment tools from the NCBI BLAST package (blastn, megablast and others) are used to align short nucleotide or amino acid sequences, like genes or proteins. For example, one may search for gene homologs in the genome of interest using blastn. The search is performed in three steps:

• exact matches (seeds or words) are searched between query and subject sequences; the word size is specified by the -word_size option,
• the seeds are expanded without using gaps (the -xdrop_ungap option),
• the obtained gap-free alignments are expanded using gaps (the -xdrop_gap and -xdrop_gap_final options).

The general rule is that the smaller is the word size, the more sensitive but also slower is the alignment search.

The third group of alignment tools is read aligners like BWA or bowtie. They are optimized to align the large number of reads and usually do well with their default parameters. The read aligners are used in comparative genomics analysis, for example, to detect genomic variants between individuals of the same or related species by aligning the reads to the genome.

As far as I understand, you have an assembled reference genome and a pool of reads from related individuals and you are to study their genomic variability. For that analysis, the read aligners seem to be the most appropriate tools. I would advise you to use bowtie2 - it is a reliable, convenient and fast tool that supports running in parallel mode. Please find the link to the bowtie2 website below.

http://bowtie-bio.sourceforge.net/bowtie2/index.shtml

Hello Bilgenur

I recently read a paper describing a similar concern, Darling et al. Their final choice was LAST (http://last.cbrc.jp/). I hope this will help.

Best,

Kyungtaek

Article PhyloSift: Phylogenetic analysis of genomes and metagenomes

Article Adaptive seeds tame genomic sequence comparison

Hi Bilgenur,

have a look at http://www.nature.com/nmeth/journal/v12/n1/full/nmeth.3176.html

This is a novel alignment tool that is up to 20000 times faster than blastx.

Best wishes,

Dominik

I used last (http://last.cbrc.jp/) to make the whole genome alignment between distant species. It is fast enough to get the result (compared to the lastz, which need a cluster to make the alignment).

However, for the paired-end reads, why not just try the populate short read aligners such like bowtie and BWA first? you can adjust the parameters to make the result acceptable.

Best,

Runsheng

I suppose you should decide what is the purpose of the alignment you are going to carry out and what datasets are used as a query and a subject. It will determine the alignment tools to be used.

Tools like LASTZ or LAST perform whole-genome alignments of large assembled genome fragments (chromosomes, scaffolds, contigs). They are optimized in the way to create so-called chained alignments, that is, alignments which consist of gap-free alignment blocks separated by large gaps. Such alignments are typically used in studying genome rearrangements, synteny blocks and homologous genome regions.

Alignment tools from the NCBI BLAST package (blastn, megablast and others) are used to align short nucleotide or amino acid sequences, like genes or proteins. For example, one may search for gene homologs in the genome of interest using blastn. The search is performed in three steps:

• exact matches (seeds or words) are searched between query and subject sequences; the word size is specified by the -word_size option,
• the seeds are expanded without using gaps (the -xdrop_ungap option),
• the obtained gap-free alignments are expanded using gaps (the -xdrop_gap and -xdrop_gap_final options).

The general rule is that the smaller is the word size, the more sensitive but also slower is the alignment search.

The third group of alignment tools is read aligners like BWA or bowtie. They are optimized to align the large number of reads and usually do well with their default parameters. The read aligners are used in comparative genomics analysis, for example, to detect genomic variants between individuals of the same or related species by aligning the reads to the genome.

As far as I understand, you have an assembled reference genome and a pool of reads from related individuals and you are to study their genomic variability. For that analysis, the read aligners seem to be the most appropriate tools. I would advise you to use bowtie2 - it is a reliable, convenient and fast tool that supports running in parallel mode. Please find the link to the bowtie2 website below.

http://bowtie-bio.sourceforge.net/bowtie2/index.shtml

First you may need to collapse paired-end read files e.g http://www.biomedcentral.com/1756-0500/5/337. I believe these files are not belong to a single species, otherwise bowtie2 or bwa or lastz. If you need faster BLAST you may try BLAT and yes megablast is less sensitive but faster. Parallel programming might be useful as well.

ublast is amazingly fast and is part of the usearch package (32 bit free, 64 bit must be purchased):

http://www.drive5.com/usearch/

I do not recommend using blat.

Discontiguous megablast is somewhat in the middle between blastn and megablast and is capable of catching distant sequences. You can try running it in parallel on multiple cores or even several computers to improve your speed. Bowtie was recommended above, but if you want to catch distantly related species it might not be so good because it was designed to map related sequences (as it was also mentioned above). I also read the article about ublast and it sounds like a good idea to improve blast speed, but I haven't tried using it personally. 

Hi you may want to try BWA or Bowtie which is quick especially for short reads. Good luck!

@Alexander Josef Probst, why would not you recommend using BLAT? Any particular reason for that?

Dear Gaik Tamazian, thanks for your detailed answer. I don't have an assembled reference genome, on the contrary my reference database consists of the pool of sequences I downloaded from NCBI, and I am working on transcriptomic datasets (aka my query-input files). I want to see what is in my input file, for that reason I am comparing it to the pool of reference sequences. I don't think Bowtie would be suitable because as I mentioned earlier I would like to catch distantly related species as well.

Hi,

You could give NextGenMap a try (http://cibiv.github.io/NextGenMap/). It is capable to align reads with an Identity of 80% or lower. If you like it also supports the usage of graphic cards to compute the alignments. This would speed up the process.

It has been already used in studies to map diverged genome and RNA seq data to a reference genome/transcriptome. 

Have a look at the hp, and if you have questions I could help you.

Best

Fritz