WinXPCNVer is tool package for calculating the Vst values between two populations probe by probe in a sliding window, which could be used to detect highly differentiated variants between populations.
WinXPCNVer is tool package for calculating the Vst values between two populations probe by probe in a sliding window, which could be used to detect highly differentiated variants between populations.
Generally, there are 2 steps if there is no significant batch effect for test (reference) population:
- Calculate Vst for each probe
calcVst.pl -t pop1.locus_summary -r pop2.locus_summary
- Calculate Vst-w (the averaged Vst of the top N probes in the sliding window of length L(bp) with at least K probes)
WinXPCNVdiffer.pl -v pop1_pop2.Vst.txt --winsize L --prb_cri K --prb_top N
The results can be ranked according to the averaged Vst-w, the window with high Vst-w is likely to be the differentiated variant between populations.
Please interpret the results with caution**!** As the microarray data could be very noisy, manually-checking the probe intensities is required and a comprehensive assessment of the signal (e.g. whether the region has been reported with CNV before, or functional annotation) is suggested.
If there is more than one batch in test (reference) population, we suggest perforing an additional step after Step1 to reduce the batch effect. This step is not required but highly suggested especially when raw data of the microarray were generated in different labs or at different time.
To illustrate this idea, we assume the test populations from two batches (named TEST1 and TEST2), and the reference population (REF) from one batch.
We can calculate Vst' as following:
Vst'=Vst[TEST1_vs_REF]+Vst[TEST2_vs_REF]-Vst[TEST1_vs_TEST2]
In this case, we have a script Vst_prime_3pop.pl to calculate the above formular (See Vst_prime_3pop.pl for detail).
The basic idea is to increase the true difference of the variance between populations and reduce the noise caused by batch effect.
As an extension, a test population contains i batches and a reference population contains j batches.
The Vst' can be calculated as following:
Vst'=SUM(pairwise_Vst[TEST_vs_REF])-SUM(pairwise_Vst[TEST_vs_TEST])-SUM(pairwise_Vst[REF_vs_REF])
We provide a simple calculator Vst_prime_calculator.pl to add or minus Vst values from two Vst files.
After this additional step, we can use the Vst' as input for calculating Vst-w.
calcVst.pl is used to calculate basic Vst value for each marker/probeset between two populations (denoted as test population and reference population) in the script.
The typical input files are the '.locus_summary' from Birdsuite[1] results, however, it is not restricted to the Birdsuite output. Any files following the format could be used as input:
The first four columns are unique markerid/probesetid, chromosome, position, and probe type (not important for the calculation), followed by the individual intensity values.
For example:
| marker_id | chrom | pos | type | sample1 | sample2 | sample3 |
|---|---|---|---|---|---|---|
| CN_134 | 3 | 7689933 | C | 1111.34 | 1345.95 | 976.53 |
| rs1234 | 5 | 3234455 | S | 9999 | 7777 | 8888 |
Please note that the header 'marker_id chrom pos type sample1 sample2 sample3 …' is required in the input file.
The leading annotation lines starting with '#' will be automatically ignored.
The markers/probesets and their order should be exactly the same in the two input files.
====================calcVst.pl USAGE====================
calcVst.pl
--tpop_lsum|-t test.pop.locus_summary
the locus_summary of the test population
--rpop_lsum|-r reference.pop.locus_summary
the locus_summary of the reference population
--help|-h
print help document
# the relationship of the test and reference populations here is relative,
this only determines the name of the output file (i.e., test.pop_reference.pop.Vst.txt ).
============================================================
The output file is the Vst values for markers/probesets. The first four columns are the same as that of the input .locus_summary files. The fifth column is the Vst value. Note that we now only calculate for autosomes.
Vst_prime_calculator.pl is a simple calculator (supporting operations only '+' and '-') for two Vst files.
The format of the input Vst file is the same as the output of the calcVst.pl (namely, markerid/probesetid -> chrom -> position -> probe_type -> Vst).
Note that the markers/probesets and their order should be exactly the same in the two input files.
====================Vst_prime_calculator.pl USAGE====================
Vst_prime_calculator.pl
--vst_1|-1 vst_1.txt
the first file of vst between two populations
--vst_2|-2 vst_2.txt
the second file of vst between two populations
--op|o {a|add|m|minus|+|-}
specify the operation to apply to the two vst files. only 'add', 'a', 'minus',
'm', '+', or '-' are allowed.
--help|-h
print help document
# Algorithm: Vst' = Vst_1 +/- Vst_2
============================================================
The output file is the so-called Vst'. The file format is also the same as the output of calcVst.pl.
Vst_prime_3pop.pl is a toy script to perform the method to reduce the noise for the same population genotyped from different batches, which is described in our paper, based on three Vst files.
Note that the markers/probesets and their order should be exactly the same in the input files.
====================Vst_prime_3pop.pl USAGE====================
Vst_prime_3pop.pl
--vst_1|-1 rpop_t1pop.vst.txt
the vst between ref population and test1 population (namely, batch.1)
--vst_2|-2 rpop_t2pop.vst.txt
the vst between ref population and test2 population (namely, batch.2)
--vst_3|-3 t1pop_t2pop.vst.txt
the vst between test1 population (namely, batch.1) and test2 population (namely,
batch.2)
--help|-h
print help document
# Algorithm: Vst' = Vst_r_t1 + Vst_r_t2 - Vst_t1_t2
============================================================
The output file is the so-called Vst'. The file format is also the same as the output of calcVst.pl.
WinXPCNVdiffer.pl performs a window-based scanning on Vst signals. Given a number of markers/probesets as a criterion (--prb_cri), if a window includes markers/probesets satisfying the criterion, the top n (--prb_top) Vst values are averaged as the Vst-w value for this window.
Note that the input file (Vst or Vst') should be sorted first by chromosome, and then by position.
====================WinXPCNVdiffer.pl USAGE====================
WinXPCNVdiffer.pl
--vst_prime|-v vst_prime.txt or Vst.txt
the vst_prime or Vst results (from 'Vst_prime_calculator.pl', or 'calcVst.pl',
respectively)
--winsize integer
size of the sliding window, in basepair, 1000 as default.
--prb_cri integer
the criterion of the number of markers/probesets in a window, 3 as default.
--prb_top integer
the number of the top Vst values in a window to calculate the Vst-w statistics.
--overlap number
the overlap of sliding windows, could be less than 1 (means proportion of the
winsize) or be an integer (means in basepairs), automatically detected.
0 as default.
--help|-h
print help document
============================================================
The output includes 7 columns, with one window per line. The first three columns give the region of the window (i.e., chromosome, boundary marker/probeset positions for start and end), the fourth column is the number of the markers/probesets in this window, followed by the mean and sd of the top n Vst values in this window (i.e., Vst-w statistics), and the last column gives the markerids/probesetids.
Note that we now only calculate for autosomes.
We give a simple example to illustrate the output of the filenames of our scripts.
For example, we have two '.locus_summary' files: 'pop1.locus_summary' and 'pop2.locus_summary'. Then we run commands (in '>' lines) and give the NEW output files (in '#' lines), supposing users in Linux/Unix environment:
>ls
# pop1.locus_summary pop2.locus_summary
>calcVst.pl -t pop1.locus_summary -r pop2.locus_summary
# pop1_pop2.Vst.txt
>calcVst.pl -t pop2.locus_summary -r pop1.locus_summary
# pop2_pop1.Vst.txt
>Vst_prime_calculator.pl -1 pop1_pop2.Vst.txt -2 pop2_pop1.Vst.txt -o '+'
# pop1_pop2.Vst.pop2_pop1.Vst.add.Vst_prime.txt
>Vst_prime_calculator.pl -1 pop1_pop2.Vst.txt -2 pop2_pop1.Vst.txt -o m
# pop1_pop2.Vst.pop2_pop1.Vst.minus.Vst_prime.txt
>WinXPCNVdiffer.pl -v pop1_pop2.Vst.txt --winsize 3000 --prb_cri 5 --prb_top 3
# pop1_pop2.Vst.VstW.win3000.5prb.top3.txt
When using WinXPCNVer, please cite:
Lou H, Lu Y, Lu D, Fu R, Wang X, Feng Q, Wu S, Yang Y, Li S, Kang L, Guan Y, Hoh BP, Chung YJ, Jin L, Su B, Xu S. A 3.4-kb Copy-Number Deletion near EPAS1 Is Significantly Enriched in High-Altitude Tibetans but Absent from the Denisovan Sequence. Am J Hum Genet. 2015 Jul 2;97(1):54-66. doi: 10.1016/j.ajhg.2015.05.005. Epub 2015 Jun 11. PMID: 26073780; PMCID: PMC4572470.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent