Page 2 of 10 FirstFirst 1234 ... LastLast
Results 11 to 20 of 95

Thread: Updated BAM Analysis Kit. Any interest?

  1. #11
    Registered Users
    Posts
    2,740
    Sex

    Quote Originally Posted by teepean47 View Post
    Can you post any examples of biased calls?
    Felix has posted a large number of samples online. You can compare them to the academic releases, like from the Reich Lab, to see the differences.

  2. #12
    Registered Users
    Posts
    31
    Sex
    Nationality
    Finnish
    Y-DNA
    R1b-Z142

    Quote Originally Posted by Generalissimo View Post
    Felix's BAM analysis kit produces biased calls, so that samples analyzed with this setup aren't compatible with the academically released genotypes, so I'm not sure what the excitement is about?

    Fix this problem and then yeah, I'll be excited as well.
    Quote Originally Posted by Generalissimo View Post
    Felix has posted a large number of samples online. You can compare them to the academic releases, like from the Reich Lab, to see the differences.
    One example, please, so I don't have to go through several of them.

  3. The Following User Says Thank You to teepean47 For This Useful Post:

     rothaer (12-04-2018)

  4. #13
    Registered Users
    Posts
    2,740
    Sex

    Quote Originally Posted by teepean47 View Post
    One example, please, so I don't have to go through several of them.
    I can't be bothered. I deleted all of the files from Felix's site from my system.

    You're missing my point anyway. The point is that there's a flaw in the BAM analysis kit, and it's either caused by trying to make diploid calls where there's not enough data to do so, and/or by this problem...

    Important Note: You should definitely use the -B flag, which disables base alignment quality recalibration. This mechanism is turned on by default and causes huge reference bias with low coverage ancient DNA data. This flag disables the mechanism.
    https://github.com/stschiff/sequenceTools

  5. #14
    Registered Users
    Posts
    31
    Sex
    Nationality
    Finnish
    Y-DNA
    R1b-Z142

    Quote Originally Posted by Generalissimo View Post
    I can't be bothered. I deleted all of the files from Felix's site from my system.

    You're missing my point anyway. The point is that there's a flaw in the BAM analysis kit, and it's either caused by trying to make diploid calls where there's not enough data to do so, and/or by this problem...
    It is really difficult for me to start verifying your observations without any data to compare the output of BAM Analysis Kit with.

  6. #15
    Registered Users
    Posts
    31
    Sex
    Nationality
    Finnish
    Y-DNA
    R1b-Z142

    I am looking into the BAQ issue but in the meantime here's BAM Analysis Kit 2.0. GitHub wasn't working for me today so here is a Google Drive link. As I am not allowed to post link, please replace hxxps with https. Please let me know if you have any issues!

    Short readme:

    After downloading and extracting the archive run BAM Analysis Kit.exe and choose a small test file. The GUI will detect your CPU and memory and adjust settings in bamkit.config after selecting Start Analysis. I have included a separate .bat file called native_bam.bat that uses samtools compiled with MinGW. Picard will give warnings about new syntax and GATK about log4j-core. You can dismiss those warnings.

    hxxps://drive.google.com/open?id=1M6bPqkP0kXVrpGrVLOPXocHw16lw9uuD

    Small changelog:

    *Cygwin updated to the latest version

    *Partial Intel GKL support. Multi-threaded PairHMM is not supported as I have not been able to get OpenMP working with MinGW-w64. Compiled with Clang 7.0.0 as GCC's AVX support on Windows is flaky. AVX-512 instructions might be supported but I cannot test it as my CPU does not support them.

    hxxps://github.com/Intel-HLS/GKL

    *Picard updated to the latest version and now includes Intel inflate/deflate from GKL

    *GATK updated to 3.8-1. Includes routines from Intel GKL apart from multi-threaded PairHMM.

    *lobSTR updated to 4.0.6

    *Samtools updated to latest version from upstream because htslib got a patch recently that has improved speed.

    *Java updated to version 8

    *I have included HaploGrep to give a more detailed mtDNA haplogroup. See out\mtDNA-haplogroup.txt

  7. The Following User Says Thank You to teepean47 For This Useful Post:

     lukaszM (11-03-2018)

  8. #16
    Registered Users
    Posts
    2,740
    Sex

    Quote Originally Posted by teepean47 View Post
    It is really difficult for me to start verifying your observations without any data to compare the output of BAM Analysis Kit with.
    Download some samples from the Reich Lab, and see how they compare to your versions. Not very difficult.

    If you can't go through such a basic step to validate your pipeline output, then don't bother with it.

    Of course you're aware that calling accurate diploid genotypes from low quality (even less than 20x coverage) data is very difficult, hence, in academic works low coverage data are called as pseudo-haploid genotypes.

    Here's an example of one way around that. But if you don't have a way around it, then there's no point using your pipeline.

    This version of the program will take any genotype with a low quality heterozygote call (Q<30) and convert to the next best homozygote call.
    https://github.com/kveeramah/aDNA_GenoCaller

  9. The Following User Says Thank You to Generalissimo For This Useful Post:

     khanabadoshi (11-04-2018)

  10. #17
    Registered Users
    Posts
    31
    Sex
    Nationality
    Finnish
    Y-DNA
    R1b-Z142

    Quote Originally Posted by Generalissimo View Post
    I can't be bothered. I deleted all of the files from Felix's site from my system.

    You're missing my point anyway. The point is that there's a flaw in the BAM analysis kit, and it's either caused by trying to make diploid calls where there's not enough data to do so, and/or by this problem...
    I have been reading samtools' source and BAQ is not used the operations BAM Analysis Kit does. GATK does the realignment and BAQ is switched off by default and is not enabled in the kit.

  11. #18
    Banned
    Posts
    6,337
    Sex
    Location
    Torun
    Ethnicity
    Central 75% + 25% Mazovia
    Nationality
    Pole
    Y-DNA
    R1a > M198 > YP 1337

    Poland
    edited
    Last edited by lukaszM; 11-04-2018 at 11:27 PM.

  12. #19
    Registered Users
    Posts
    2,740
    Sex

    Quote Originally Posted by teepean47 View Post
    I have been reading samtools' source and BAQ is not used the operations BAM Analysis Kit does. GATK does the realignment and BAQ is switched off by default and is not enabled in the kit.
    Produce a PLINK dataset with a couple of ancient samples that are also available at the Reich Lab, like the Iranian farmers or whatever, and I'll run some formal stats comparing your samples with the Reich Lab samples, and then we'll see.

  13. #20
    Registered Users
    Posts
    31
    Sex
    Nationality
    Finnish
    Y-DNA
    R1b-Z142

    Quote Originally Posted by Generalissimo View Post
    Produce a PLINK dataset with a couple of ancient samples that are also available at the Reich Lab, like the Iranian farmers or whatever, and I'll run some formal stats comparing your samples with the Reich Lab samples, and then we'll see.
    Sure. Do you have any preference in output format or options or do you want whatever plink outputs?

Page 2 of 10 FirstFirst 1234 ... LastLast

Similar Threads

  1. 23andme YDNA Hg/SNP of Interest
    By asquecco in forum 23andMe
    Replies: 3
    Last Post: 11-13-2018, 09:15 PM
  2. Replies: 51
    Last Post: 09-03-2017, 10:55 PM
  3. South Asian? This may be of interest.
    By Amerijoe in forum Other
    Replies: 4
    Last Post: 07-17-2017, 04:23 AM
  4. What motivates your interest?
    By A Norfolk L-M20 in forum General
    Replies: 48
    Last Post: 10-13-2016, 02:17 AM
  5. Replies: 2
    Last Post: 04-16-2015, 01:29 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •