Rna

[skly]

I aligned RNA sequenceing data to human genome using HISAT. To my surprised, the alignment rates were too low, about 75%. So, I want to know:
1) Why was the alignment rate (~75%) so low?
2) As for the RNA sequencing data of human tissue, which of alignment software is suitable between HISAT, TopHat2 and STAR?

Thanks a lot!

[dpryan]

No one can tell you why you only got a 75% alignment rate without looking at your data. Perhaps you have some notable adapter contamination or quality issues and didn't trim (or used end to end alignment). Perhaps you have bacterial or other contamination, have a look at some of the unaligned reads.

Generally HISAT or STAR would be used for RNAseq data, since tophat2 is just too slow. I should note that STAR is nice in that it will give you a summary of why it couldn't align reads (e.g., they're too short, too many possible hits, etc.).

[skly]

No one can tell you why you only got a 75% alignment rate without looking at your data. Perhaps you have some notable adapter contamination or quality issues and didn't trim (or used end to end alignment). Perhaps you have bacterial or other contamination, have a look at some of the unaligned reads.

Generally HISAT or STAR would be used for RNAseq data, since tophat2 is just too slow. I should note that STAR is nice in that it will give you a summary of why it couldn't align reads (e.g., they're too short, too many possible hits, etc.).

Thank you, dpryan！
Actually, these RNA sequencing data have be removed the adapter and be filtered low quality reads.
I suspected that HISAT caused the low mapping rate, firstly. So, for the same reference genome and the same input RNA sequencing data, I run tophat2 (v2.1.0). But the mapping rate of tophat2 is about 70%. So, I think that the low mapping rate maybe is not the result of alignment software, but reference genome I chose.
It is worth mentioning that the reference genome sequence downloaded from Ensembl is “dna_rm” type (masked genomic DNA). Maybe it is the cause. Next, I will test the “dna” type (unmasked) reference genome.

[dpryan]

Yup, using a hard masked reference would cause that. Just use either the soft masked or unmasked (the results will be the same).

[skly]

Yup, using a hard masked reference would cause that. Just use either the soft masked or unmasked (the results will be the same).

Thank you, dpryan. I replaced the masked reference with unmasked genome reference. The mapping rate was about 90%, but I found one another question. The discordant alignment rate was so high. The following are links. Could you help me to see the results? Thank you so much.

http://seqanswers.com/forums/showthread.php?t=62304

Ps: The HISAT results of soft masked reference were as same as the unmasked. And the parameters were defaults.