Blast experts, please help - (Dec/13/2010 )

Hi,
When I did blast search for some proteins I had strange output results. I got higher score for a sequence with less identity and less gaps to the query. As you can see below Ad46 score was 215 and Ad28 score was 214 although Ad28 is 95% identical with only one gap. I have found same results with several proteins. So can somebody help me to understand what is going on?
Thanks



>gb|AAX70942.1| 130R
Length=130

Score = 215 bits (547), Expect = 1e-54, Method: Compositional matrix adjust.
Identities = 112/130 (87%), Positives = 121/130 (94%), Gaps = 2/130 (1%)

Query 1 MQTLLILLSLFSHTLAD--DYSKCKFVELWNFLDCYDAKMDMPSYYLVIVGIVMVCSCTF 58
MQTLL+L+SL S LA DYSKCKF ++WNFLDCY K+DMPSYYLVIVG+VMVCSCTF
Sbjct 1 MQTLLMLISLLSPALATSADYSKCKFADIWNFLDCYQEKIDMPSYYLVIVGVVMVCSCTF 60

Query 59 FAIMIYPCFDLGWNSVEAFTYTLESSSLASTPPPTPPPRRNQFPLIQYLEEPPPRPPSTV 118
FAIMIYPCF+LGWNSVEAFTYTLE+SSLASTPPPTPPPRRNQFP+IQYLEEPPPRPPSTV
Sbjct 61 FAIMIYPCFNLGWNSVEAFTYTLENSSLASTPPPTPPPRRNQFPMIQYLEEPPPRPPSTV 120

Query 119 SYFHITGGDD 128
SYFHITGGDD
Sbjct 121 SYFHITGGDD 130


>gb|ACQ91169.1| E3 14.5K-like protein
Length=129

Score = 214 bits (544), Expect = 3e-54, Method: Compositional matrix adjust.
Identities = 122/129 (95%), Positives = 125/129 (97%), Gaps = 1/129 (0%)

Query 1 MQTLLILLSLFSHTLAD-DYSKCKFVELWNFLDCYDAKMDMPSYYLVIVGIVMVCSCTFF 59
MQTLL+LLSL S +AD DYSKCKFVELWNFLDCYDAKMDMPSYYLVIVGIVMVCSCTFF
Sbjct 1 MQTLLLLLSLLSPAIADPDYSKCKFVELWNFLDCYDAKMDMPSYYLVIVGIVMVCSCTFF 60

Query 60 AIMIYPCFDLGWNSVEAFTYTLESSSLASTPPPTPPPRRNQFPLIQYLEEPPPRPPSTVS 119
AIMIYPCFDLGWNSVEAFTYTLESSSLASTPPPTPPPRRNQFP+IQYLEEPPPRPPSTVS
Sbjct 61 AIMIYPCFDLGWNSVEAFTYTLESSSLASTPPPTPPPRRNQFPMIQYLEEPPPRPPSTVS 120

Query 120 YFHITGGDD 128
YFHITGGDD
Sbjct 121 YFHITGGDD 129

The problem was sorted out by turning the Compositional adjustments off from algorithm parameters