For paired end libraries, how do I know which is forward and which is reverse?

[skmotay]

For paired end libraries, how do I know which is forward and which is reverse?

I've got the files in galaxy, but I don't know which are forward and which are reverse, just that they're paired.

[GenoMax]

File with R1 in the name is the first read and R2 is the second read (other end) from the same set of fragments.

[skmotay]

The file names don't have those. After the genotype, they just say lane_4 of lane_5

[Brian Bushnell]

Can you post a complete filename, and also the first 8 lines from that file?

[GenoMax]

Have the reads been processed in any way (i.e. collapsed into a single long read, possible if the insert is short and reads overlap)?

If reads ARE really paired, then they can't be in separate lanes (there would be two files per lane).

[skmotay]

File name:

Galaxy3-[WT_2_lane4.fastq].fastq (then the next file is the same, except lane5. each genotype has two files, one lane four and one lane 5)

@GWZHISEQ02:321YMKACXX:5:1101:1297:1992 1:N:0:ATCACG
TTTGGATCGTTCACAAATTGAGAGAGGTTTTCGCGTAGGAAGTACTTGNCC
+
CCCFFFFFHHHFHJJIIJJJJJJJJJJ@GHJIIJJJGIJEIJBFHIJI###
@GWZHISEQ02:321YMKACXX:5:1101:1692:1983 1:N:0:ATCACG
CCCGCCCTTCTTGCCGCGCGGCNGCATCATTCGGTTGGGCACGCCGCNNNN
+
@@@DDADDHFFHHIII@GGIIG#07<FDHIIG@F5AGHFEBE??A@B####
@GWZHISEQ02:321YMKACXX:5:1101:2408:1992 1:N:0:ATCACG
GTACTTGTAGAGCTCGGTGTATGTAATCCTGTCCTCAGCGGTCCAGCTNAA
+
@<<DDEDDHGFHHIDEHCGIIGIDHHJ@HHIGIAFEIGEEGIIIGHIG###
@GWZHISEQ02:321YMKACXX:5:1101:3073:1994 1:N:0:ATCACG
CCGAGTTCTCCTGGTACCAGCAGATCGAGATGGGCTACGACCCCCAGCNGA
+
@@CFFDFFHHHHHJGHIJJIIJJJJJJJJIIJJIGIHIIJJJJJJJJJ###
@GWZHISEQ02:321YMKACXX:5:1101:3748:1991 1:N:0:ATCACG
GTGCGTTCGCTGGAGAATGTGTGCCGCGACCTGATCAACGGTGCAAAGNAC
+
@@@DDDDDAHFDFFHFEFHIIJEGGIEGEHIJGBGIGGIIG?EHHFDF###
@GWZHISEQ02:321YMKACXX:5:1101:3653:1998 1:N:0:ATCACG
CTGACTGGCATGATCCAGTCGACGATGATGGCCATGCCCTACATGGACAAA
+
CCCFFFFFHHHHHJJJJJJJJJJJHJJJJJJJJJJJJJJIJJJJJJJJJJ#
@GWZHISEQ02:321YMKACXX:5:1101:5002:1996 1:N:0:ATCACG
CCTCAGTTCTCCTGGTTCCACCAGATCGAGATGGGCTACGATCCACAANTG
+
???;:=2=A=D<DEBE::+<<F3<33A?1811111)11:6)0000?D####
@GWZHISEQ02:321YMKACXX:5:1101:5470:1986 1:N:0:ATCACG
CTGGATATCAATAATGCTCTCCNTAGGGATATTTCCCGCAAATTTGANNNN
+
CCCFFFFFHHHHHJJJJJJJJJ#3AGIJJJJJJJJJJJJJJJJJJJJ####



When I try to take the two with the same geno# and map them as paired end, I get this:
Left reads:
Input: 22670964
Mapped: 21232534 (93.7% of input)
of these: 1499440 ( 7.1%) have multiple alignments (0 have >20)
Right reads:
Input: 22516885
Mapped: 21088604 (93.7% of input)
of these: 1487965 ( 7.1%) have multiple alignments (0 have >20)
93.7% overall read alignment rate.

Aligned pairs: 19856733
of these: 99084 ( 0.5%) have multiple alignments
and: 18762256 (94.5%) are discordant alignments
4.9% concordant pair alignment rate.

[Brian Bushnell]

These reads are not paired.

@GWZHISEQ02:321YMKACXX:5:1101:1297:1992 1:N:0:ATCACG

That 1 indicates it is read 1. So you get a high discordant rate because you are mapping unrelated single-ended reads as if they were paired.

[GenoMax]

If your description is right then they are not paired end reads. Just a sample run in two lanes.

For them to be true paired-end reads, each sample will need to have two files (from the same lane) with the naming convention described here: http://en.wikipedia.org/wiki/FASTQ_f...ce_identifiers

[skmotay]

Well, that explains...a whole lot. And eases as much frustration!

(If only the data generator had left some notes or would return an email...)

I thought biological reps were unnecessary for RNA seq?

[GenoMax]

I thought biological reps were unnecessary for RNA seq?

It is true that technical replicates are unnecessary but biological replicates are a must if you hope to get any meaningful results.

[skmotay]

Yup, sorry I meant technical reps are unnecessary.