search

Xinglab / TideHunter

TideHunter: efficient and sensitive tandem repeat detection from noisy long reads using seed-and-chain
https://github.com/yangao07/TideHunter
MIT License
19 stars 2 forks source link

Illegal instruction (core dumped) #21

Open fgponce opened 1 month ago

fgponce commented 1 month ago

Hi, I love that you have made this tool. installed using conda and I ran this command "TideHunter -u barcode18_test.fastq > units3.fa" and get a file full of the split out subreads/repeats. However, I'm now keen to look at the consensuses it can make so I removed the -u to "TideHunter barcode18_test.fastq > units3.fa" I get "Illegal instruction (core dumped)"

I've tried to make consensus from different files and changed the output name and type eg -f 2 or -f 3 but get the same error...

Any ideas why it runs but can't make the consensus etc?

I updated it using conda but still the same problem.

Rob

yangao07 commented 1 month ago

I think this is a known issue when TideHunter is run on a relatively old machine. Can you try to re-install it from the source?

fgponce commented 1 month ago

OK. Bugger its on our new HCS cluster hahaha. Do you think its my conda environment that is too old and/or the server set up/hardware?

yangao07 commented 1 month ago

I am not sure about your case, Again have you tried installing it from source? If that works, it means the machine is old.

fgponce commented 1 month ago

Cheers. We installed it with the makefile and had to play a bout a bit but finally got the right sub-modules etc and it ran.

So thanks for the top tip.

My stuff is actually inverted repeats separated by a hairpin. Do you have advice about getting the repeats/consensus from there? It's a bit odd since it'll go S-InsertAS-HPAS-InsertAS-HP_S-InsertAS-HPetc

So I guess the software will try make a "S-InsertAS-HPAS-Insert" unit from these reads...

So I shouldn't use HP sequence as a pseudo 5' or 3' adapters...which may make it difficult to get accurate full length consensus for my inserts.

Could I force it to make units split at each HP sequence? Then take every other unit entry and make two different consensus (S and AS) and then a combine into a single consensus...or add a step to reverse complement every second insert to contribute to the consensus calling?

Hahaha sorry to ask but it seems like its great tool that's nearly exactly what I need!!

Rob

From: Yan Gao @.> Sent: Saturday, August 3, 2024 5:53 AM To: Xinglab/TideHunter @.> Cc: Robert Day @.>; Author @.> Subject: Re: [Xinglab/TideHunter] Illegal instruction (core dumped) (Issue #21)

I am not sure about your case, Again have you tried installing it from source? If that works, it means the machine is old.

- Reply to this email directly, view it on GitHubhttps://github.com/Xinglab/TideHunter/issues/21#issuecomment-2265875428, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABNJUUZAGLS66HXYFXQZUKLZPPBPNAVCNFSM6AAAAABLYGN6GWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENRVHA3TKNBSHA. You are receiving this because you authored the thread.Message ID: @.**@.>>