Mosaik, gigaBayes and paired-end output

[Gianza]

Hello everybody,
I'm facing a weird problem, really hope that somebody can help me.

I have aligned both 454 and Solexa paired-ends reads on a transcriptome reference for SNP mining purposes.

The two alignments have been merged with MosaikMerge, and MosaikAssembler run fine to build contigs. The MosaikAssembler module itself provide a flag to output reads and qualities (FASTA) which can be used together with ACE directly for gigabayes.
(Just note that all single fasta,qual and ace from each contig have been concatenated, and ace modified accordingly to its format).
Everything was PERFECT with only 454 alignment..confirming that I'm not messing up with concatenation.

BUT, using also paired-ends from Illumina, both ACE and FASTA file in output loose their /1 and /2 tags on read names which identify mates.
Thus, when running gigaBayes, it get confused by this, looking for one mates in ACE, finding another in FASTA. Indeed, it stop as soon unpadded read lengths don't match.

It seems absurd to me that the developer of both tools (Marth Lab) didn't considerate this aspect. Therefore I'm wishing to find my stupid-dumb-error.

the two fastq files as input for Mosaik are as following,

mates_1.fastq

@Sample1-0110:5:1:1002:3520#TGACCA/1
GCAAGACAGCAAGTGACAAATCCCCCATCGATAAATTGTGGGTCGTCTAGTGATC
+Sample1-0110:5:1:1002:3520#TGACCA/1
)-58806)*))/1111+757788707733555AAAAA8776523535AAAAA###
@sample1-0110:5:1:1003:10121#TGACCA/1
AACAATGTCATCGATTTCGCTATCGTATCTACAGAGCGCATGCTCCTATTCAAATCTTGGACCACAAAACAAAAGAG
+Sample1-0110:5:1:1003:10121#TGACCA/1
+*.1AAAAAAAAAAAAAAAAAAAAAAAAA7::7:AAAAAAAAAAAAAAAAAAA7AAAAA787:8AAAAAAA2772AA


mates_2.fastq

@Sample1-0110:5:1:1002:3520#TGACCA/2
AGGCTGCGACGAACAGGGCTCGACAAAGCATTCGAGAATATATAGTTCTTCGACGCTGCTTGATCACTAGACGACC
+Sample1-0110:5:1:1002:3520#TGACCA/2
A?A::B=BB??;?<?BB?+8?BB;BAAA.A?8+??B?BA?BABBABBBBB*?;<?AB?B95??*5AB>BBBB>98A
@Sample1-0110:5:1:1003:10121#TGACCA/2
GTGTTTAAACCAAAAGGGCTCATTGGTTTTGATTTGGAAACATTTGTGTTGTTCTCTTTTGTTTTGTGGTCCAAGA
+Sample1-0110:5:1:1003:10121#TGACCA/2
etc...

I will really appreciate any help! I'm not a real bioinformatician, thus almost impossible for me to work around this.