[webpage] Pre-build index builds are not accessible and transparent about hisat2

[zhanxiaoyu]

Hello, I've downloaded the index online too and found it useless, so I built the index by myself following the shell script they provided and successfully.
Firstly, you have to prepare SNP file, GTF file,and homo sapiens genome fa. All of them have been downloaded the same as what they used.
In the index you downloaded, you can find a file named make_grch38_snp_tran.sh, you can change the shell script a little, and pay a little attention to the version of python, and the you can build the same index that exactly the official provided.

[bgruening]

@zhanxiaoyu thanks for your detailed answer!
I found this as well, but I think this should be included in the webpage and the links should be fixed.
The documentation about building index and all sorts of recommendations are a little bit weak, imho.

[zhanxiaoyu]

Yes, I thinks so.
There is so limit information about this software.
Can you build the index successfully?

[bgruening]

I could, but only the standard index so far, without SNP. It's still queued on your Galaxy cluster :)

[zhanxiaoyu]

Why?
What's problem about your SNP?
Actually, though I build the four kinds of index successfully, it still failed in alignment.

[bgruening]

Memory, I have requested a lot :)
Yes align also crashed for me with: https://github.com/infphilo/hisat2/issues/6

[zhanxiaoyu]

I've never received signal 13 but signal 6 and 4.
I requested 6g memory to build the index with SNP.

[bgruening]

6g? In the documentation is something written from up-to 200g.

If you use --snp, --ss, and/or --exon, hisat2-build will need about 200GB RAM for the human genome size as index building involves a graph construction. 
Otherwise, you will be able to build an index on your desktop with 8GB RAM.

[zhanxiaoyu]

oh,I am so sorry to make a mistake, I requested more than 200GB indeed.
Actually, I met a new problem, which is the final alignment statistic result is not the same as the one I got from bam file.

[bgruening]

I think it would help @infphilo if you can fill all these errors under a new issue.
Seems that HISAT is not usable at the moment :(

[zhanxiaoyu]

I just hope Daehwan Kim will help us solve all of this question.
It is so tired to use HISAT2

[infphilo]

@bgruening, thank you for letting me know the problem. You can open it using http instead of https for now. We'll enable users to use both http and https sometime soon. Once you download and extract any zipped file, you will find a shell script that has instructions how to build indexes.

[infphilo]

I fixed these access issues.