Why does plink-qc.nf check for samplesheet, sample.phe, and idatpat?

[mchimenti]

Hello, I'm trying to run plink-qc.nf on my Linux HPC cluster with:

nextflow run -c my_nf.config /Users/mchiment/.nextflow/assets/h3abionet/h3agwas/plink-qc.nf -profile sgeSingularity --samplesheet "0" --idatpat ""

Where "sgeSingularity" is:
sgeSingularity {

process.executor = "sge"
singularity.autoMounts = true
singularity.enabled = true
process.queue = queue

}

I keep getting errors like this:

N E X T F L O W ~ version 19.01.0
Launching /Users/mchiment/.nextflow/assets/h3abionet/h3agwas/plink-qc.nf [agitated_lavoisier] - revision: 418f49aacb
Check idatpat=true ************** is it a valid parameter -- are you using one rather than two - signs or vice-versa
ERROR ~ Argument of file function cannot be empty

-- Check script 'plink-qc.nf' at line: 237 or see '.nextflow.log' file for more details

Line 237:

sample_sheet_ch = file(params.samplesheet)

---

Why is it checking for a samplesheet when the directions say we only need a BIM/BED/FAM file as input for this pipeline?

Similarly, if the pipeline also requires a "sample.phe" file, you should put that in the instructions. I had to create this by hand.

[mchimenti]

OK, I seem to have gotten the pipeline to at least kick off with:
nextflow run /Users/mchiment/.nextflow/assets/h3abionet/h3agwas/plink-qc.nf -profile sgeSingularity --input_pat "willour_041919_bed" --samplesheet "./input/samplesheet.csv"

Where there is a pseudo-phenotype file I created called "sample.phe" under the "input" directory (where the BED/BIM/FAM files are). It looks like this:

IID PHE
235-17-002 0
235-17-004 0
235-17-005 0
235-17-008 0
235-17-009 0
235-17-010 0
235-17-011 0
235-17-012 0
235-17-014 0
235-17-015 0
235-17-016 0
235-17-017 0
etc....

AND there is a pseudo-samplesheet.csv that I also had to create, it looks like this:
Institute Sample Label, Manifest Sex, Batch Comment
235-17-002,0,0
235-17-004,0,0
235-17-005,0,0
235-17-008,0,0
235-17-009,0,0
235-17-010,0,0
235-17-011,0,0
235-17-012,0,0
235-17-014,0,0
235-17-015,0,0
235-17-016,0,0
235-17-017,0,0
235-17-020,0,0
.....

---

So now the pipeline starts, and then fails on compPCA with this:

N E X T F L O W ~ version 19.01.0
Launching /Users/mchiment/.nextflow/assets/h3abionet/h3agwas/plink-qc.nf [silly_hopper] - revision: 418f49aacb
Institute Sample Label
Manifest Sex
Batch Comment
WARN: Access to undefined parameter idatpat -- Initialise it to a default value eg. params.idatpat = some_value
Sexinfo available command
[warm up] executor > sge
WARN: Output set must define at least two component -- Check process compPCA
Pulling Singularity image docker://quay.io/h3abionet_org/py3plink [cache /Users/mchiment/.singularity/quay.io-h3abionet_org-py3plink.img]
[b3/6994b5] Submitted process > getDuplicateMarkers (1)
[8a/e420a9] Submitted process > sampleSheet
[0e/862028] Submitted process > inMD5 (1)
[d5/e3a627] Submitted process > removeDuplicateSNPs (1)
[50/e06403] Submitted process > getInitMAF (1)
[31/58fbc3] Submitted process > identifyIndivDiscSexinfo (1)
[09/90c9f1] Submitted process > generateIndivMissingnessPlot (1)
[06/2aff3c] Submitted process > getX (1)
[3d/3d37c6] Submitted process > removeQCPhase1 (1)
[41/aad570] Submitted process > generateSnpMissingnessPlot (1)
[9a/ce5202] Submitted process > showInitMAF (1)
[1a/c430c1] Submitted process > showHWEStats (1)
ERROR ~ Error executing process > 'compPCA (1)'

Caused by:
Failed to submit process to grid scheduler for execution

Command executed:

qsub -terse .command.run
Command exit status:
1

Command output:
Unable to run job: "job" denied: use parallel environments instead of requesting slots explicitly
Exiting.

Work dir:
/Users/mchiment/nf_h3agwas/work/f8/89e74e040f2acbb1ed53ae90b0b58d

Tip: you can try to figure out what's wrong by changing to the process work dir and showing the script file named .command.sh

-- Check '.nextflow.log' file for details
WARN: Killing pending tasks (2)

---

How do I get compPCA to run? Do I change max_plink_cores to 1 in the nextflow.config?

Also, do my "pseudo" PHE and SAMPLESHEET files have the right format?

Thanks, Michael

[shaze]

Dear Michael

You've asked several questions, not sure what you've sorted out, so I'll try to deal with them in order

First, my apologies. We made some changes late Thursday to add some error checking code and managed add some bugs. I've fixed and committed the changes, so please pull the latest version. I apologise for wasting your time on this.

1. The first error you reported, the one with "idatpat", was our fault and is fixed. It should be "idpat" (and you only need to specify if you are using a samplesheet)
2. You do not need to provide a samplesheet. This is an advanced feature. The default value is "0" and you do not need to specify this. My bug fix should have made your problem go away. (As an aside the value of using the samplesheet is if your samplesheet contains per-sample call rate and GenCall 10 scores and you are using those for QC. If not, there's no point in using it).
3. With respect to phenotypes for QC: what is compulsory is a case/control value for which you expect no difference at genome-wide level (even though there may be some SNPs which do differ). This could be a genuine case/control value or something you use as a marker to check that there are no differences between groups. For example, you could use the sex column of a fam file -- one strategy would be just to copy the fam file and put a header row in. Note that the IDs have to be given in PLINK-style, with FID and IID columns. You only seem to have the IID column. I know in many cases the IID will be unique but there are cases that our pipeline has to cope with where they're not and it's hard to distinguish. To recap, what's compulsory is to define params.case_control which is a file name (with header lines) and params.case_control_col which the column name.
4. To directly answer your question : your samplesheet value looks good though without call rate and GC10 score it won't be helpful. I would suggest not setting that value. The phenotype file needs a FID column. My suggestion -- make a copy of the fam file and add a header row.
5. There are more advanced features which are optional of doing batch analysis of different sorts and require defining params.phenotype and params.batch (and the corresponding col columns). For example, in our study which is a multi-site study and each site sent their DNA in four different batches we wanted to check genome-wide differences at a site level and at a batch level. It also helps track down problems that might have occured in a particular shipment. Typically the batch, phenotype an d case_control file would all be the same file, but this need not be the case.
6. The final problem you report Unable to run job: "job" denied: use parallel environments instead of requesting slots explicitly is not specific to us but about how you need to run Nextflow under SGE. I have not done this, but I think it's well documented here https://gitter.im/nextflow-io/nextflow/archives/2015/12/28 and in the link to the Nextflow documentation.

Here is a very simple example config file, where ``batch4-mod.phe` is the fam file with a header tow and a column header "CC" in the sixth column

params.input_dir = "./"
params.input_pat = "batch4"
params.output = "b4-qc"
params.output_dir = "qc"

params.case_control_col="CC"
params.case_control="batch4-mod.phe"

Please let us know if this helps or if you have any other problems. and we'd appreciate any other feedback you have

[mchimenti]

Dear Scott,

Thank you for your detailed and timely reply and for supporting this pipeline. I have made some progress but I am still running into issues at the "batch_proc.py" step.

I did a fresh pull (as of today 5/28/19) of the pipeline. I am running with:

nextflow run /Users/mchiment/.nextflow/assets/h3abionet/h3agwas/plink-qc.nf -profile sgeSingularity

In the nextflow.config I specified:

case_control         = "${params.input_dir}/sample.phe"
case_control_col     = "PHENO"

head ./input/sample.phe
FID IID PID MID SEX PHENO
1 235-17-005 0 0 2 1
2 235-17-043 0 0 1 1
3 235-17-501 0 0 2 0
4 235-17-002 0 0 2 1
5 235-17-526 0 0 2 0
6 235-17-045 0 0 1 1
7 235-17-012 0 0 2 1
8 NIMH_Control_B_009 0 0 0 0
9 235-17-008 0 0 1 1

*********Launching nextflow gives:

N E X T F L O W ~ version 19.01.0
Launching /Users/mchiment/.nextflow/assets/h3abionet/h3agwas/plink-qc.nf [crazy_albattani] - revision: 3ee82ddf72
WARN: Access to undefined parameter samplesheet -- Initialise it to a default value eg. params.samplesheet = some_value
The file is called ******
0
Sexinfo available command
[warm up] executor > sge
WARN: Output set must define at least two component -- Check process compPCA
[7a/bd7122] Submitted process > inMD5 (1)
[41/a1520e] Submitted process > sampleSheet
[de/550568] Submitted process > getDuplicateMarkers (1)
[10/dc628a] Submitted process > removeDuplicateSNPs (1)
[ed/dd3f38] Submitted process > getInitMAF (1)
[2e/47eade] Submitted process > removeQCPhase1 (1)
[f5/a54594] Submitted process > getX (1)
[5c/62de14] Submitted process > identifyIndivDiscSexinfo (1)
[cc/c5df10] Submitted process > generateIndivMissingnessPlot (1)
[97/0d084e] Submitted process > generateSnpMissingnessPlot (1)
[b7/089a58] Submitted process > compPCA (1)
[1a/d652b4] Submitted process > calculateSampleHeterozygosity (1)
[4e/5d019e] Submitted process > showInitMAF (1)
[b0/07e6a7] Submitted process > pruneForIBD (1)
[d3/974e8a] Submitted process > analyseX (1)
[f3/89194b] Submitted process > showHWEStats (1)
[3b/64a0ec] Submitted process > getBadIndivsMissingHet (1)
[a7/615aa4] Submitted process > generateMissHetPlot (1)
[96/d19e71] Submitted process > findRelatedIndiv (1)
[47/a182be] Submitted process > drawPCA (1)
[68/bf88ca] Submitted process > removeQCIndivs (1)
[f8/d1d7ce] Submitted process > batchProc (1)
Command exit status:
1

Command output:
(empty)

Command error:
Traceback (most recent call last):
File "/usr/lib/python3.6/site-packages/pandas/core/indexes/base.py", line 2525, in get_loc
return self._engine.get_loc(key)
File "pandas/_libs/index.pyx", line 117, in pandas._libs.index.IndexEngine.get_loc
File "pandas/_libs/index.pyx", line 139, in pandas._libs.index.IndexEngine.get_loc
File "pandas/_libs/hashtable_class_helper.pxi", line 1265, in pandas._libs.hashtable.PyObjectHashTable.get_item
File "pandas/_libs/hashtable_class_helper.pxi", line 1273, in pandas._libs.hashtable.PyObjectHashTable.get_item
KeyError: 'FID'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File ".command.sh", line 614, in
figs = plotPCs(base,eigs,pfrm,args.pheno_col,bfrm,args.batch_col)
File ".command.sh", line 237, in plotPCs
df=df.merge(batch,on=['FID','IID'],how='inner')
File "/usr/lib/python3.6/site-packages/pandas/core/frame.py", line 5370, in merge
copy=copy, indicator=indicator, validate=validate)
File "/usr/lib/python3.6/site-packages/pandas/core/reshape/merge.py", line 57, in merge
validate=validate)
File "/usr/lib/python3.6/site-packages/pandas/core/reshape/merge.py", line 565, in init
self.join_names) = self._get_merge_keys()
File "/usr/lib/python3.6/site-packages/pandas/core/reshape/merge.py", line 824, in _get_merge_keys
right_keys.append(right[rk]._values)
File "/usr/lib/python3.6/site-packages/pandas/core/frame.py", line 2139, in getitem
return self._getitem_column(key)
File "/usr/lib/python3.6/site-packages/pandas/core/frame.py", line 2146, in _getitem_column
return self._get_item_cache(key)
File "/usr/lib/python3.6/site-packages/pandas/core/generic.py", line 1842, in _get_item_cache
values = self._data.get(item)
File "/usr/lib/python3.6/site-packages/pandas/core/internals.py", line 3843, in get
loc = self.items.get_loc(item)
File "/usr/lib/python3.6/site-packages/pandas/core/indexes/base.py", line 2527, in get_loc
return self._engine.get_loc(self._maybe_cast_indexer(key))
File "pandas/_libs/index.pyx", line 117, in pandas._libs.index.IndexEngine.get_loc
File "pandas/_libs/index.pyx", line 139, in pandas._libs.index.IndexEngine.get_loc
File "pandas/_libs/hashtable_class_helper.pxi", line 1265, in pandas._libs.hashtable.PyObjectHashTable.get_item
File "pandas/_libs/hashtable_class_helper.pxi", line 1273, in pandas._libs.hashtable.PyObjectHashTable.get_item
KeyError: 'FID'

Work dir:
/Users/mchiment/nf_h3agwas/work/f8/d1d7ce1317369bc218bae701257dbb

Tip: you can replicate the issue by changing to the process work dir and entering the command bash .command.run

-- Check '.nextflow.log' file for details

---

[shaze]

Could I see the nextflow.config file please?

Scott

[mchimenti]

mchiment@argon-login-2: h3agwas$ cat nextflow.config

py3Image = "quay.io/h3abionet_org/py3plink"
gemmaImage="quay.io/h3abionet_org/h3agwas-gemma"
latexImage="quay.io/h3abionet_org/h3agwas-texlive"
swarmPort = '2376'
queue = 'IIHG'

manifest {
homePage = 'http://github.com/h3abionet/h3agwas'
description = 'GWAS Pipeline for H3Africa'
mainScript = 'plink-qc.nf'
}

aws {
accessKey =''
secretKey =''
region ='eu-west-1'
}

cloud {
imageId = "ami-710b9108" // specify your AMI id here
instanceType = "m4.xlarge"
subnetId = "null"
sharedStorageId = "null"
sharedStorageMount = "/mnt/shared"
bootStorageSize = "20GB" // Size of disk for images spawned
// instanceStorageMount = "" // Set a common mount point for images
// instanceStorageDevice = "" // Set a common block device for images
autoscale {
enabled = true
maxInstances = 1
terminateWhenIdle = true
}

}

params {

// Directories
work_dir                = "/$PWD"
input_dir               = "${params.work_dir}/input"
output_dir              = "${params.work_dir}/output"
scripts                 = "${params.work_dir}/scripts"
output                  = "out"


max_forks            = 95
// Data
input_pat            = "willour_041919_bed"


high_ld_regions_fname = ""
sexinfo_available    = true
cut_het_high         = 0.343
cut_het_low           = 0.15
cut_diff_miss         = "0.05"
cut_maf                = "0.01"
cut_mind              = "0.02"
cut_geno              = 0.01
cut_hwe               = 0.008
pi_hat                = 0.11
super_pi_hat 	  = 0.7
f_lo_male             = 0.8 // default for F-sex check -- >= means male
f_hi_female           = 0.2 //  <= means female
gc10                 = 0.4  // 10% Gen Call confidence -- 0.4 is very low
case_control         = "${params.input_dir}/sample.phe"
case_control_col     = "PHENO"


phenotype = 0
pheno_col = "all"
batch = 0
batch_col = 0

samplesize           = 0  // How many individuals in each genotype report 0=ALL
strandreport         =""
manifest             =  ""
idpat                =  0  // or "(\\w+)-DNA_(\\w+)_.*" or ".*_(.*)"


// Needed for topbottom.nf -- uncomment and put in details
// reference        = ""
// output_align     = ""
// samplesheet      = ""
// chipdescription  = ""

accessKey            = ""
secretKey            = ""
region               = "eu-west-1"
AMI                  = "ami-710b9108"
instanceType         = "m4.xlarge"
bootStorageSize      = "20GB"
maxInstances         = "1"

plink_mem_req        = "750MB"
other_mem_req        = "750MB"
big_time             = '12h'
sharedStorageMount   = "/mnt/shared"
max_plink_cores      = 4

}
profiles {

// For execution on a local machine, no containerization. -- Default
standard {
    process.executor = 'local'
}

slurm {
process.executor = 'slurm'
process.queue = queue
}

// Execute pipeline with Docker locally
docker {
    process.executor = 'local'
    docker.remove      = true
    docker.runOptions  = '--rm'
    docker.registry    = 'quay.io'
    docker.enabled     = true
    docker.temp        = 'auto'
    docker.fixOwnership= true
    docker.process.executor = 'local'
}


// Execute pipeline with Docker Swarm setup
dockerSwarm {
    docker.remove      = true
    docker.runOptions  = '--rm'
    docker.registry    = 'quay.io'
    docker.enabled     = true
    docker.temp        = 'auto'
    docker.fixOwnership= true
    docker.process.executor = 'local'
    docker.engineOptions = "-H :$swarmPort"
}

// For execution on a PBS scheduler, no containerization.
pbs {
    process.executor = 'pbs'
    process.queue = queue
}


// For execution on a PBS scheduler with containerization.
pbsDocker {

    process.executor = 'pbs'
    docker.remove      = true
    docker.runOptions  = '--rm'
    docker.registry    = 'quay.io'
    docker.enabled     = true
    docker.temp        = 'auto'
    docker.fixOwnership= true
}


// For execution on a SLURM scheduler, no containerization.
slurm {
    process.executor = 'slurm'
    process.queue = queue
}


// For execution on a PBS scheduler with containerisation.
slurmDocker {
    process.executor = 'slurm'
    docker.remove      = true
    docker.runOptions  = '--rm'
    docker.registry    = 'quay.io'
    docker.enabled     = true
    docker.temp        = 'auto'
    docker.fixOwnership= true
}


singularity.cacheDir = "${HOME}/.singularity"

singularity {
    singularity.autoMounts = true
    singularity.enabled = true
    process.executor = 'local'
 }

// For execution on a SLURM scheduler with singularity
slurmSingularity {

    process.executor = 'slurm'
    singularity.autoMounts = true
    singularity.enabled = true
    process.queue = queue

}

sgeSingularity {

process.executor = "sge"
singularity.autoMounts = true
singularity.enabled = true
process.queue = queue
penv = "smp"

}
slurmDocker {
    process.executor = 'slurm'
    process.queue = queue
    docker.remove      = true
    docker.runOptions  = '--rm'
    docker.registry    = 'quay.io'
    docker.enabled     = true
    docker.temp        = 'auto'
    docker.fixOwnership= true

}

}

process {
container = py3Image

withLabel:bigMem {
  memory = '8GB'
}

withLabel: gemma {
      container = gemmaImage
}

}

timeline {
enabled=true
file = "nextflow_reports/timeline.html"
}

report {
enabled = true
file = "nextflow_reports/report.html"
}

[shaze]

Very odd, will have to look at it tomorrow.

What’s odd about this is that as far as I can see, params.batch = 0, in the nextflow.config file. This being the case the code that crashes should not be called.

Will try to recreate the problem. What OS are you running?

Scott

[shaze]

Could you please email me the file /Users/mchiment/nf_h3agwas/work/f8/d1d7ce1317369bc218bae701257dbb/.command.sh scott.hazelhurst@wits.ac.za

[mchimenti]

Very odd, will have to look at it tomorrow.

What’s odd about this is that as far as I can see, params.batch = 0, in the nextflow.config file. This being the case the code that crashes should not be called.

Will try to recreate the problem. What OS are you running?

Scott

This is what I thought too. You said batch analysis was off by default, and I did not change that parameter in the config file.

[shaze]

Michael,

As reported privately I was able to reproduce the problem which is due to differences in the way different versions of pandas handles certain operations. I have now made changes that should make the scripts robust to the version of pandas and have tested on pandas 22 (which I was able to reproduce the problem) and pandas 24. I have pushed through changes. Please try now.

[mchimenti]

Hi Scott, thank you I will test it now. Just a suggestion, but you could simply pull a docker/singularity container with python/pandas in it and use that to run your custom scripts. That way you'd always have control over pandas versions.

[shaze]

Thanks for the suggestion

We do have these containers and I should be doing regression testing against these to make sure I am always backwards compatible.

We don't have complete control because some of the environments we are designing for do not suppot docker/singularity and so users will install their own versions. We can't check against everythiing of course, but ideall what I shoud be doing is in my own environment testing against latest releases of software and then always doing regression testing against our own docker images, which is what I had failed to do.

[mchimenti]

Hi Scott,
So the pipeline does now run past the batch proc step. There is a new error:

ERROR ~ Error executing process > 'calculateSnpSkewStatus (1)'

Caused by:
Missing output file(s) willour_041919_bed-nd-c-c.missing expected by process calculateSnpSkewStatus (1)

Command executed:

cp sample.phe cc.phe
plink --threads 4 --autosome --bfile willour_041919_bed-nd-c-c --pheno cc.phe --pheno-name PHENO --test-missing mperm=10000 --hardy --out willour_041919_bed-nd-c-c
if ! [ -e willour_041919_bed-nd-c-c.missing.mperm ]; then
echo " CHR SNP EMP1 EMP2 " > willour_041919_bed-nd-c-c.missing.mperm
fi

Command exit status:
0

Command output:
PLINK v1.90b5.2 64-bit (9 Jan 2018) www.cog-genomics.org/plink/1.9/
(C) 2005-2018 Shaun Purcell, Christopher Chang GNU General Public License v3
Logging to willour_041919_bed-nd-c-c.log.
Options in effect:
--autosome
--bfile willour_041919_bed-nd-c-c
--hardy
--out willour_041919_bed-nd-c-c
--pheno cc.phe
--pheno-name PHENO
--test-missing mperm=10000
--threads 4

257523 MB RAM detected; reserving 128761 MB for main workspace.
281985 variants loaded from .bim file.
44 people (13 males, 31 females) loaded from .fam.
27 phenotype values present after --pheno.
Using up to 4 threads (change this with --threads).
Before main variant filters, 44 founders and 0 nonfounders present.
Calculating allele frequencies... done.
--hardy: Writing Hardy-Weinberg report (founders only) to
willour_041919_bed-nd-c-c.hwe ... done.
281985 variants and 44 people pass filters and QC.
Among remaining phenotypes, 0 are cases and 27 are controls. (17 phenotypes
are missing.)

The pipeline is assigning all of the "case" phenotypes to missing, which is causing this step to fail.

[mchimenti]

mchiment@argon-login-2: output$ head willour_041919_bed-nd-c.fam
1 235-17-005 0 0 2 1
2 235-17-043 0 0 1 1
3 235-17-501 0 0 2 -9
4 235-17-002 0 0 2 1
5 235-17-526 0 0 2 -9
6 235-17-045 0 0 1 1
7 235-17-012 0 0 2 1
8 NIMH_Control_B_009 0 0 0 -9
9 235-17-008 0 0 1 1
10 235-17-523 0 0 2 -9

---

Actually, it looks like all of the controls are being assigned as missing. I'm not sure I understand what's going on here.

There are supposed to be 27 cases and 17 controls, + 4 internal controls.

[mchimenti]

Here is cc.phe from nf_h3agwas/work/46/efa0c2bfa91d7cab51875a96f0af31 which is just sample.phe copied from the inputs.

This is correct. So now this is confusing, why is PLINK getting confused and thinking that cases are controls and controls are missing?

cat cc.phe
FID IID PID MID SEX PHENO
1 235-17-005 0 0 2 1
2 235-17-043 0 0 1 1
3 235-17-501 0 0 2 0
4 235-17-002 0 0 2 1
5 235-17-526 0 0 2 0
6 235-17-045 0 0 1 1
7 235-17-012 0 0 2 1
8 NIMH_Control_B_009 0 0 0 0
9 235-17-008 0 0 1 1
10 235-17-523 0 0 2 0
11 235-17-011 0 0 2 1
12 235-17-530 0 0 1 0
13 NIMH_Control_A_009 0 0 0 0
14 235-17-503 0 0 2 0
15 235-17-065 0 0 2 1
16 235-17-506 0 0 1 0
17 235-17-063 0 0 2 1
18 235-17-521 0 0 1 0
19 235-17-060 0 0 2 1
20 235-17-009 0 0 2 1
21 235-17-017 0 0 2 1
22 235-17-535 0 0 1 0
23 235-17-058 0 0 2 1
24 235-17-051 0 0 1 1
25 NIMH_Control_A_011 0 0 0 0
26 235-17-508 0 0 1 0
27 235-17-046 0 0 2 1
28 235-17-512 0 0 2 0
29 235-17-015 0 0 1 1
30 235-17-042 0 0 2 1
31 235-17-050 0 0 2 1
32 235-17-520 0 0 2 0
33 NIMH_Control_B_011 0 0 0 0
34 235-17-004 0 0 2 1
35 235-17-047 0 0 2 1
36 235-17-537 0 0 2 0
37 235-17-020 0 0 1 1
38 235-17-525 0 0 1 0
39 235-17-522 0 0 2 0
40 235-17-066 0 0 2 1
41 235-17-511 0 0 2 0
42 235-17-529 0 0 2 0
43 235-17-010 0 0 2 1
44 235-17-040 0 0 2 1
45 235-17-014 0 0 1 1
46 235-17-016 0 0 2 1
47 235-17-536 0 0 2 0
48 235-17-024 0 0 2 1

[shaze]

Dear Michael

Gosh, you're giving our pipeline a thorough testing!

Here the issue is that PLINK uses 1 for cases and 2 for controls. Both 0 and -9 mean missing values.

See https://www.cog-genomics.org/plink/1.9/formats#fam for more detail.

Of course, our pipeline should fail gracefully, so I wil definitely put this on the list of things to improve

Regards

Scott