Bowtie/1.1.2 Encountered reference seq with gaps

[Sow]

Hello,

Can anyone decipher this code for me..

} else if(cat >= 2) {
if(lc != -1 && off == 0) off++;
lc = -1;
off++; // skip over gap character and increment
} else if(c == '>') {
if(off > 0 && lastc == '>') {
cerr << "Warning: Encountered reference sequence with only gaps" << endl;
} else if(lastc == '>') {
cerr << "Warning: Encountered empty reference sequence" << endl;
}
lastc = '>';
//return RefRecord(off, 0, false);
return RefRecord((uint32_t)off, 0, first);
}
c = in.get();
if(c == -1) {
// End-of-file
if(off > 0 && lastc == '>') {
cerr << "Warning: Encountered reference sequence with only gaps" << endl;
} else if(lastc == '>') {
cerr << "Warning: Encountered empty reference sequence" << endl;
}
lastc = -1;
//return RefRecord(off, 0, false);
return RefRecord((uint32_t)off, 0, first);
}
}
assert(!rparms.color || (lc != -1));
assert_eq(1, asc2dnacat[c]); // C must be unambiguous base
if(off > 0 && rparms.color && first) {
// Handle the case where the first record has ambiguous
// characters but we're in color space; one of those counts is
// spurious
off--;
}

I'm trying to build a bowtie index on a genome file but it returns the warning
"Encountered reference sequence with gaps".

I'm not sure what that means. My genome is not Rep Masked to have N (at least that I know of). And when I do downstream aligning only 20% of my reads aligns with the genome. When I created an index for an older version of the genome I did not encounter these warnings and I had more reads aligning to the genome (more than 50%). I am limited to using the new version of the genome since I am using a pipeline and the gff file I use later in the pipeline have the same headers as the new version of the genome.

My question is
1. What does the warning "encountered....." mean and how do I solve the issue?
2. Can the low percentage of reads aligning to the genome be because of these warnings?

Thanks in advance for your help!