As others have noted, if all you have for now is the raw sequence data, i.e. files with millions of lines, each line having a short sequence of 75-100 base pairs, then for the first step you will need access to a complete exome analysis pipeline for short read alignment, variant calling and variant annotation. There are commercial programs (CLC Bio, SoftGenetics) that are more user friendly but very expensive and require high end PC workstations. If you do not have these then you will have to find a bioinformaticist to collaborate, probably with a promise of co-authorship. It is almost impossible to figure out how to do all that if you are completely on your own, unless you are a real bioinformaticist yourself. That is all JUST the first step. The primary analysis pipeline will generate around 15,000 potential variants for each individual sample, that have to be filtered using various quality score metrics and minor allele frequencies in public databases - none of this being user friendly either. At the very end, if you do everything correctly, you should have for each individual sample a set of 300-500 rare variants in coding exons or adjacent splice sites (excluding synonymous variants). They will be mostly heterozygous, and mostly in different genes. So the final step is to look for shared genes or shared specific variants among 12 such exomes, each with 3-500 different genes and variants. A bioinformaticist who can do all the previous steps can look for such shared genes/variants pretty easily, although it takes some programming skills. As others have noted, depending on whether your samples are from related individuals, that can help. If they are just 12 unrelated individuals with the same medical phenotype, then it's simply a matter of looking for sharing. Finally, it is realistic to expect that all of this will take you 6 months to a year to do the analysis, learn how to interpret the results, and identify potential candidates. Good luck!

HI,

We need some more information in order to help.

1) What type of data do you have? Sanger sequence, exome, whole genome ... raw data or aligned data ?

2) How many people do you have sequence for?

3) Do you have a specific gene of interest or is this genome-wide?

Thanks

thank you dear scott

I have exome data

Two goups each have 6 people

I inspect whole exome, not spesific gene.

multiple sequence alignment may be a good method to find conserved pattern in genes of diseased and healthy person..

but you have to first sort your whole genome sequence in groups of possible genes which are affected during diseased condition...

and then analyze the whole subsets to find the possible mutation sites..

That's a pretty big question. The basic process is read QC, alignment, variant calling, variant annotation, variant comparisons between groups and then validation. But you might also consider copy number changes as possible drivers of disease. If you're new to exome analysis, then good place to start is with the Galaxy Project:

https://usegalaxy.org/

It has lots of point and click tools to help you achieve the above steps. Let me know if you need help building a workflow. But be careful putting patient data online.

Also, bear in mind that you're likely to find thousands of variants in each individual and identifying the causal variant(s) is definitely not straightforward.

If your individuals are related you could use haplotypes to establish the chromosomes phase , then you can use Linkage analysis. I believe JMP Genomics will do something like that from NGS data. On the other hand even "healthy" controls will show somewhere between 1000s and 10000s SNP,some of them disease associated (depending on your stringency). 

Last try this site for your VCF analysis (free): http://www.gene-talk.de/ 

Dear Artur; After determining candidate genes for 12 people, Mutation regions will be checked for more people.

Dear Huseyin,

The Exome data analysis for prediction of sites of mutations which causes disease among groups of population. you can try A robust method or Biotool such as Snv PRioritization via the INtegration of Genomic data (SPRING), is a bioinformatics method for the prediction of disease-causing nonsynonymous single nucleotide variants (SNVs) in exome sequencing studies.

A stand alone software also available on

http://bioinfo.au.tsinghua.edu.cn/spring/software.php

The SPRING software (Release 13.12.01) for Windows (both 32-bit and 64-bit) is available free of charge for academic usage.

First analyze the two different dataset to identify causative SNP present only in diseased samples. This approach will help with small number of samples to identify a rare allele variant associated with the disease. Then the validation of association of variant with disease can be performed based on the SNP location in the human genome

I assume you got FastQ data, so you have to process them through an NGS pipeline, as already stated. However, in order to define the best pipe, you should probably provide some more data: are the case/controls related? trios maybe? is the disease hereditary/de novo? do you know if it is dominant/recessive? sex linked? this information is very important in order to choose the best strategy.

As others have noted, if all you have for now is the raw sequence data, i.e. files with millions of lines, each line having a short sequence of 75-100 base pairs, then for the first step you will need access to a complete exome analysis pipeline for short read alignment, variant calling and variant annotation. There are commercial programs (CLC Bio, SoftGenetics) that are more user friendly but very expensive and require high end PC workstations. If you do not have these then you will have to find a bioinformaticist to collaborate, probably with a promise of co-authorship. It is almost impossible to figure out how to do all that if you are completely on your own, unless you are a real bioinformaticist yourself. That is all JUST the first step. The primary analysis pipeline will generate around 15,000 potential variants for each individual sample, that have to be filtered using various quality score metrics and minor allele frequencies in public databases - none of this being user friendly either. At the very end, if you do everything correctly, you should have for each individual sample a set of 300-500 rare variants in coding exons or adjacent splice sites (excluding synonymous variants). They will be mostly heterozygous, and mostly in different genes. So the final step is to look for shared genes or shared specific variants among 12 such exomes, each with 3-500 different genes and variants. A bioinformaticist who can do all the previous steps can look for such shared genes/variants pretty easily, although it takes some programming skills. As others have noted, depending on whether your samples are from related individuals, that can help. If they are just 12 unrelated individuals with the same medical phenotype, then it's simply a matter of looking for sharing. Finally, it is realistic to expect that all of this will take you 6 months to a year to do the analysis, learn how to interpret the results, and identify potential candidates. Good luck!

You can use SNPsift software to do this analysis within groups. Its a free software

Thank you for all valuable comments.

It all depends to the disorder you study. If it is a rare monogenic disorder with high penetrance and for example recessive pattern of inheritance then you have a good chance of finding the causal gene with the number of cases you have without requiring any control, but if you study a complex genetic disorder like cancer, I have to say that your chance of finding the causal mutation with this number of cases and controls you have would be very low.

Generally speaking, the analysis of the NGS data specially the tertiary analysis which involves finding your candidate causal mutations among the about 24K variants you have in each exome data could be very complicated and you would be better off to collaborate with people who has experience in this field rather than wanting to do it yourself if you do not have any experience on it. If you like, my lab can help you on this.

I completely agree with the work flow suggested by Mark, if you have access to a Computational biologist with experience in NGS analysis you will know if your data supports your hypothesis or you need more samples as suggested by Mohammad. if your intention is to play with data yourself then there are couple of options e.g. Galaxy server, there you have look at some of the pipelines for DNA-seq or you can build your own. That will at least you give the variant calls in each sample. Second part of your analysis case/control comparison can again be done using few open source tools or commercial ones like Golden Helix, Ingenuity, etc. Please note all these s/w comes with free trial, you can use that period to analyze, annotate, prioritize your variants. Perhaps you might like the ease of use and buy it for your future projects.  

Thanks a lot for all answers. I try to coping with this problem through yours suggestions. I might need your help in future.

If you are looking for an insertion/deletion mutation, try using mutation surveyor software, the trial version can process up to 20 or so samples at a time. You need to specify a reference sequence and align all the other test sequences against the reference sequence. This software is capable of deconvoluting the sequence got after direct sequencing and the presence of an indel will be shown as a sift in nucleotides.

THANK YOU VERY MUCH DEAR LİNDA. I WILL TRY IT. IF I HAVE CONFUSING QUESTION CAN I ASK YOU.