If I'm understanding the specs in a comment and your regex, perhaps the following will be helpful:

use strict;
use warnings;

# Remove (exclude) lines beginning with "bl", "BL" or "_" and releases
+ with a label with xx.yy.zzz.nnn(n)
while (<DATA>) {
    print unless /^(?:bl|_)|\.\d{3}\.\d{3,}$/i;
}

__DATA__
_STUDYABCD1234_1.00
_STUDYABCD1234_1.00.5678
STUDYABCD1234_1.00
STUDYABCD1234_1.00.000
p_STUDYABCD1234_1.00.000
p_STUDYABCD1234_1.00.000.5678
bl_STUDYABCD1234_1.00.000
bl_STUDYABCD1234_1.00.000.5678
BL_STUDYABCD1234_1.00.000
BL_STUDYABCD1234_1.00.000.5678
[download]

Output:

STUDYABCD1234_1.00
STUDYABCD1234_1.00.000
p_STUDYABCD1234_1.00.000
[download]

The regex:

/^(?:bl|_)|\.\d{3}\.\d{3,}$/i
 ^     ^  ^ ^ ^    ^ ^    ^ ^       
 |     |  | | |    | |    | |       
 |     |  | | |    | |    | + - Case-insensitive
 |     |  | | |    | |    + - From the end of the string
 |     |  | | |    | + - Three or more digits
 |     |  | | |    + - A decimal point
 |     |  | | + - Three digits
 |     |  | + - A decimal point
 |     |  + - OR
 |     + - Match either "bl" or "_"
 + - From the beginning of the string
[download]

The script just skips (excludes) those lines that you don't want, instead of making them blank.