Convert a paired SRA file to FASTA file

Question

[m.koohi.m]

I want to convert a paired SRA file to FASTA file to do some downstream analysis. For example the following link that is a metagenomics wgs paired sequencing: http://www.ncbi.nlm.nih.gov/sra/?term=ERR1018199 I know I can use fastq-dump and I did same as bellow:

./fastq-dump myfile.sra --fasta

and it gives me:

>ERR1018199.1 FCD0JMKACXX:8:1101:1051:3172#ACTTGAAT length=200
GGAAGCGGTGTTCTGTTTCTCCTTCCATATATTTCCACAGACTGGCATATTCTTCTTTCTTCTCGCCCAT
TTCCTGCAATTGCAGCATATATTGCCAGATGATGAAAAAAATATACATCAAGGAAAACAAGTCTATCTGG
CAATATATGCTGCAATTGCAGGAAATGGGCGAGAAGAAAGAAGAATATGCCAGTCTGTGG
>ERR1018199.2 FCD0JMKACXX:8:1101:1348:3168#ACTTGAAT length=200
CTTGCAGATTCTACAAAAAGAGTGTTTCATAAACTGGTCTATCAAAAGAAAGGTTAAACTCAGTGAGTTG
AACCCACACATCACAAAGTAGCTTCTGAGATATGTGGGTAATATCTGTATGGATGTTTGTATGATTGATG
TTACTGACATTGATTGCAAAGAAGGCGACAGCGTTGAGATTTTCGGAGATCATCTTCCTA
>ERR1018199.3 FCD0JMKACXX:8:1101:1451:3174#ACTTGAAT length=199
GTATCATGACCGGTCGTTCGGGCAACAACATTTGGTGTATCAGTCCGATGTTCGACCTCAACAAACCGAC

and if I use --split-files option. It gives me:

ERR1018199_1.fasta:

>ERR1018199.1 FCD0JMKACXX:8:1101:1051:3172#ACTTGAAT length=100
GGAAGCGGTGTTCTGTTTCTCCTTCCATATATTTCCACAGACTGGCATATTCTTCTTTCTTCTCGCCCAT
TTCCTGCAATTGCAGCATATATTGCCAGAT
>ERR1018199.2 FCD0JMKACXX:8:1101:1348:3168#ACTTGAAT length=100
CTTGCAGATTCTACAAAAAGAGTGTTTCATAAACTGGTCTATCAAAAGAAAGGTTAAACTCAGTGAGTTG
AACCCACACATCACAAAGTAGCTTCTGAGA
>ERR1018199.3 FCD0JMKACXX:8:1101:1451:3174#ACTTGAAT length=100
GTATCATGACCGGTCGTTCGGGCAACAACATTTGGTGTATCAGTCCGATGTTCGACCTCAACAAACCGAC
TATTTTACTAAATTCCCACATCGATACCGT
>ERR1018199.4 FCD0JMKACXX:8:1101:1254:3223#ACTTGAAT length=100

ERR1018199_2.fasta:

>ERR1018199.1 FCD0JMKACXX:8:1101:1051:3172#ACTTGAAT length=100
GATGAAAAAAATATACATCAAGGAAAACAAGTCTATCTGGCAATATATGCTGCAATTGCAGGAAATGGGC
GAGAAGAAAGAAGAATATGCCAGTCTGTGG
>ERR1018199.2 FCD0JMKACXX:8:1101:1348:3168#ACTTGAAT length=100
TATGTGGGTAATATCTGTATGGATGTTTGTATGATTGATGTTACTGACATTGATTGCAAAGAAGGCGACA
GCGTTGAGATTTTCGGAGATCATCTTCCTA
>ERR1018199.3 FCD0JMKACXX:8:1101:1451:3174#ACTTGAAT length=99
GGGAGGTAGTTGGGGCAATACGCTTTCTATTCCGTTTTTTCCGGAAACTTCTTCTTCACATGAGGCAAGA
AAGATCAGATTGTATGCCTGCTCGGTTAT
>ERR1018199.4 FCD0JMKACXX:8:1101:1254:3223#ACTTGAAT length=100

My questions is, the first command simply concatenate the two reads with same read number. is it true? Can I use that for further analysis (for example assemble) or I have to use --split-files to produce two different files and apply my pipeline on those files?

Or more simpler can we say that we have the following sequence in our (meta)genome? or not?

>ERR1018199.1 FCD0JMKACXX:8:1101:1051:3172#ACTTGAAT length=200
GGAAGCGGTGTTCTGTTTCTCCTTCCATATATTTCCACAGACTGGCATATTCTTCTTTCTTCTCGCCCAT
TTCCTGCAATTGCAGCATATATTGCCAGATGATGAAAAAAATATACATCAAGGAAAACAAGTCTATCTGG
CAATATATGCTGCAATTGCAGGAAATGGGCGAGAAGAAAGAAGAATATGCCAGTCTGTGG

Answer 1

[Istvan Albert]

The concatenated file is mainly for convenience, for example to move the data around as a single file. In general it cannot be used for downstream analysis.

There is meaning for the data in each pair and most tools will be unable to split the data unless it comes from different files.