Starting my first research project, but don't know what steps to take.

[prs321]

So basically what I am doing is looking at 20 specimens (Serratia marcescens), which is 40 fastq files and I am trying to compare them to the ancestral genome aka reference genome.

My first task was the clean up the raw data, but I am clueless. I was told to first find out whether or not each file contains good reads or bad reads (using FASTQC), then remove the adapter sequences with cutadapt or a python script. I've been tinkering with the FASTQC software and I understand the results to a certain extent, but I still don't know how to clean data that I judge to be 'bad'.

I cannot get cutadapt to work and I am not sure how I would go about using a python script because my knowledge relevant to biology is very rusty. I was told if I can't get cutadapt to work, then a python script would be faster.

After cleaning the data, my next task is gene mapping. I think I was told to use BWA and the software is installed, but I am still clueless as how I should use it.

And the final stage is analysis I think...



Can anyone point me in the right direction? I am having a difficult time understanding everything that is going on, especially trying to understand both the biology and computer science side.

Thanks.


Edit:

I am expected to know how to do all of this in Linux terminal. I have the basics down pretty much, but I have to apply it.

[mastal]

Starting my first research project, but don't know what steps to take.

You have a steep learning curve ahead, but we've all been there.

Some reading material which should help:

Basic Linux tutorial:

http://www.ee.surrey.ac.uk/Teaching/Unix/

BWA web pages:

http://bio-bwa.sourceforge.net/

Samtools (for manipulating the alignment files) web pages:

http://samtools.sourceforge.net/

Nature Methods supplement on NGS data analysis

http://www.nature.com/nmeth/journal/v6/n11s/index.html

Hope this helps,
Maria

[krobison]

BWA, as with many Linux programs, will give you some tips if you execute it with no arguments. BWA follows the model in which the first argument is a command to bwa; executing this with no further arguments will further explain that command

Is your data paired end or single end? Platform?

[Heisman]

I haven't used BWA; does someone who has know if it will perform adapter trimming when aligning? I'd be surprised if it didn't.

[westerman]

I haven't used BWA; does someone who has know if it will perform adapter trimming when aligning? I'd be surprised if it didn't.

I've always used BWA (and Bowtie2) with adapter trimmed sequences. Re-reading the manual does not indicate the BWA does adapter trimming. Quality trimming, sure, that is easy to implement. Bowtie2 also does not do adapter trimming. Given your "I'd be surprised" statement it seems like you expect aligners to do adapter trimming. Which aligner do you commonly use?

[prs321]

BWA, as with many Linux programs, will give you some tips if you execute it with no arguments. BWA follows the model in which the first argument is a command to bwa; executing this with no further arguments will further explain that command

Is your data paired end or single end? Platform?

They are paired end.

[krobison]

I haven't used BWA; does someone who has know if it will perform adapter trimming when aligning? I'd be surprised if it didn't.

BWA doesn't explicitly adapter trim (search against an adapter database), but the CIGAR for the alignments will include soft trimming (S) operations -- so this can be used to identify & remove adapters.

[jgibbons1]

Here is my typical pipeline:

(1) To clean data
a. Trim Galore to trim by quality AND remove adapter sequence

http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/

b. Remove duplicate reads with fastxtoolkit (FASTQ Collapser script)

http://hannonlab.cshl.edu/fastx_toolkit/

(2) Map reads
I've had very good luck with BOWTIE2 but use BWA if you are comfortable with it. BOWTIE2 has 4 built in parameter presents that trade off between speed and sensitivity (very fast, fast, sensitive, and very sensitive).

http://bowtie-bio.sourceforge.net/bowtie2/index.shtml

My bit of advice (from my own experiences) would be to test your pipeline with a small data set (say 1,000 reads). Once you are comfortable with that, have played around with parameters and understand how the software works and the types of output you get, you'll be ready to ramp up.

Good luck!

[Heisman]

I've always used BWA (and Bowtie2) with adapter trimmed sequences. Re-reading the manual does not indicate the BWA does adapter trimming. Quality trimming, sure, that is easy to implement. Bowtie2 also does not do adapter trimming. Given your "I'd be surprised" statement it seems like you expect aligners to do adapter trimming. Which aligner do you commonly use?

Ah, darn it. I was spoiled with Novoalign and assumed it was a more standard thing, but clearly I was wrong.