--salti is very time consuming

[yangwukaidi]

Dear Michael DeGiorgio, Ph.D. Thank you very much for your reply. As you said, the New software saltiLASSI works really well. I found --salti to be very time-consuming. I want to know a .vcf dataset of 650,000 loci from 30 individuals, and how long does it take to use --salti? The commands I use are as follows： lassip --spectra Z.chr12.lassip.hap.spectra.gz --salti --out Z.chr12 Looking forward to your reply. Best wishs for you. Yang

[szpiech]

Hi,

You can try increasing winstep size, this should help quite a bit. I saw in your other email you had window of 51 and step of 6; perhaps try increasing step to 25 or 50.

Zachary

Le jeu. 30 nov. 2023 à 3:55 PM, yangwukaidi @.***> a écrit :

Dear Michael DeGiorgio, Ph.D. Thank you very much for your reply. As you said, the New software saltiLASSI works really well. I found --salti to be very time-consuming. I want to know a .vcf dataset of 650,000 loci from 30 individuals, and how long does it take to use --salti? The commands I use are as follows： lassip --spectra Z.chr12.lassip.hap.spectra.gz --salti --out Z.chr12 Looking forward to your reply. Best wishs for you. Yang

— Reply to this email directly, view it on GitHub https://github.com/szpiech/lassip/issues/4, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABAKRQTMCGREJ5K7ODV7F43YHCM5PAVCNFSM6AAAAABABKLN46VHI2DSMVQWIX3LMV43ASLTON2WKOZSGAYTQOBUGYZDONA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[yangwukaidi]

Hi, Dear Zachary A. Szpiech Ph.D Thank you very much for your reply. As you said, increasing winstep size did reduce the time consuming, the software is running normally. When I look at the result file for --salti(win51 step51), the m value is always 0, as follows, However, this is not the case in the result file of -lassi(win51 step6), as follows, Is it normal that m is always equal to 0? What's the point of m being equal to 0? Or how can I try to improve it? Isn't this dataset only suitable for use -- lassi? Looking forward to your reply. Best. Yang

[szpiech]

Hello,

This is strange that you would get all 0. Can you tell me which version of the software you are using?

Zachary

Le ven. 1 déc. 2023 à 9:16 AM, yangwukaidi @.***> a écrit :

Hi, Dear Zachary A. Szpiech Ph.D Thank you very much for your reply. As you said, increasing winstep size did reduce the time consuming, the software is running normally. When I look at the result file for --salti(win51 step51), the m value is always 0, as follows, 1701418495825.png (view on web) https://github.com/szpiech/lassip/assets/127917612/8b1bd02d-a04b-4f6a-a6b7-3bcfbdeac9a7 However, this is not the case in the result file of -lassi(win51 step6), as follows, 1701418495820.png (view on web) https://github.com/szpiech/lassip/assets/127917612/5c701c93-50e6-4c31-8ecb-92f9c5246d4b Is it normal that m is always equal to 0? What's the point of m being equal to 0? Or how can I try to improve it? Isn't this dataset only suitable for use -- lassi? Looking forward to your reply. Best. Yang

— Reply to this email directly, view it on GitHub https://github.com/szpiech/lassip/issues/4#issuecomment-1835661098, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABAKRQQ5Z7BKHOLPFEH2TO3YHGG4HAVCNFSM6AAAAABABKLN46VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMZVGY3DCMBZHA . You are receiving this because you commented.Message ID: @.***>

[yangwukaidi]

Hi. The version I'm using is lassip v1.1.2.

[yangwukaidi]

phase1Z12.vcf.gz Z.chr12.lassip.hap.spectra.gz Z.chr12.lassip.hap.stats.gz Z.ids.pop.txt Z.salti.chr12.lassip.hap.out.gz All the relevant documents are attached, please check. Yang.

[szpiech]

Hello,

I am currently travelling for work and won't be able to look carefully at this until at least a week from now. I'm wondering if the bug fix I introduced in 1.1.2 actually created another one. I restored the v1.1.1 linux binary on the salti branch, which you can find here https://github.com/szpiech/lassip/tree/salti/bin/linux . You can try this one and let me know if the issue still exists in the meantime.

Zachary

On Fri, Dec 1, 2023 at 6:18 AM yangwukaidi @.***> wrote:

phase1Z12.vcf.gz https://github.com/szpiech/lassip/files/13525560/phase1Z12.vcf.gz Z.chr12.lassip.hap.spectra.gz https://github.com/szpiech/lassip/files/13525582/Z.chr12.lassip.hap.spectra.gz Z.chr12.lassip.hap.stats.gz https://github.com/szpiech/lassip/files/13525584/Z.chr12.lassip.hap.stats.gz Z.ids.pop.txt https://github.com/szpiech/lassip/files/13525611/Z.ids.pop.txt Z.salti.chr12.lassip.hap.out.gz https://github.com/szpiech/lassip/files/13525759/Z.salti.chr12.lassip.hap.out.gz All the relevant documents are attached, please check. Yang.

— Reply to this email directly, view it on GitHub https://github.com/szpiech/lassip/issues/4#issuecomment-1835915810, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABAKRQXL5CDA5F566MWQJXLYHG4I3AVCNFSM6AAAAABABKLN46VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMZVHEYTKOBRGA . You are receiving this because you commented.Message ID: @.***>

[yangwukaidi]

Hi, Dear Zachary A. Szpiech Ph.D You're so helpful! I tested v1.1.1 using my test data and the results were normal. So, I am currently running my complete data, I'll ask you if I have any questions. Best. Yang

[cgdmkns]

Dear @szpiech, Since I'm facing the same issue as @yangwukaidi, I'm adding my comment here. I have a salti result for 780snp window with 100 step size, all m values are 0. I'm using v1.1.1. The commands I used:(used parallel to estimate for each chromosome) lassip --vcf chr1.vcf --hapstats --salti --calc-spec --winsize 780 --winstep 100 --out Chr1_bal --pop my_popID.txt lassip --spectra Chr1_bal.lassip.hap.spectra.gz --salti --out chr1 Here is a part of the result.

Could it be related with win size?

[szpiech]

Hello,

780 does seem quite large, but I’m concerned you have a version of the program with the same bug as the first user. Would you try grabbing the mossy recent version and trying that?

Zachary

Le jeu. 4 juil. 2024 à 04:19, cgdmkns @.***> a écrit :

Dear @szpiech https://github.com/szpiech, Since I'm facing the same issue as @yangwukaidi https://github.com/yangwukaidi, I'm adding my comment here. I have a salti result for 780snp window with 100 step size, all m values are 0. I'm using v1.1.1. The commands I used:(used parallel to estimate for each chromosome) lassip --vcf chr1.vcf --hapstats --salti --calc-spec --winsize 780 --winstep 100 --out Chr1_bal --pop my_popID.txt lassip --spectra Chr1_bal.lassip.hap.spectra.gz --salti --out chr1 Here is a part of the result. saltilassi_result.xlsx https://github.com/user-attachments/files/16095490/saltilassi_result.xlsx Could it be related with win size?

— Reply to this email directly, view it on GitHub https://github.com/szpiech/lassip/issues/4#issuecomment-2208385690, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABAKRQRKL4VCMPRGIJSNFC3ZKUARZAVCNFSM6AAAAABKLBEYJGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMBYGM4DKNRZGA . You are receiving this because you were mentioned.Message ID: @.***>

[cgdmkns]

Sure, I'll do. 780 window size is based on the interval over which LD decayed below one-third of its original value relative to pairs of loci separated by 1 kb as stated in Harris and DeGiorgio,2020. I estimated the SNP density at that interval. But I'll give it a try for smaller windows as well. Thanks!

[cgdmkns]

Dear @szpiech,

It worked with the v1.2.0.Now I have m values different than 0. I appreciate your help!

Cigdem

[szpiech]

Great, glad to hear it!

Le mer. 10 juil. 2024 à 04:15, cgdmkns @.***> a écrit :

Dear @szpiech https://github.com/szpiech,

It worked with the v1.2.0.Now I have m values different than 0. I appreciate your help!

Cigdem

— Reply to this email directly, view it on GitHub https://github.com/szpiech/lassip/issues/4#issuecomment-2219846610, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABAKRQXPAAHHMAD2P4S3X53ZLTUQ5AVCNFSM6AAAAABKLBEYJGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMJZHA2DMNRRGA . You are receiving this because you were mentioned.Message ID: @.***>