특정 질병관련 유전자셋에서 다른 유전자들에 비해 더 큰 증감을 보이는 유전자가 있을때 bias 표시 (note 컬럼 참조)가 됩니다.
- 원문 참고 : The Bias Term is the product of the dataset bias and/or the bias of target molecules involved in a particular function/regulator annotation. A biased dataset is one where there is more up- than down-regulated genes or vice versa. The bias term is not considered for the activation z-score but is considered in the bias-corrected z-score (which is another column also hidden by default and can be included by selecting 'customize table' and selecting to add the bias adjusted z-score). The dataset bias is constant for any given analysis and the disease/function/regulator bias is unique for each disease/function/regulator. An example would be a disease in which all the molecules associated with this disease in literature are described to be upregulated in the disease state. Therefore the value of the bias term is representative of the disease/function/regulator bias if your dataset is not 'biased'.
이렇게 편중된 유전자 때문에 disease/function이 증가한다고 뜰 수 있으므로 주의하라는 뜻입니다.
이럴때는 신뢰성이 떨어지긴 하지만 0.05 이하 p-value를 보이는지도 함께 봐야합니다.
Default로 인해 숨겨진 column인 "bias-corrected z-score"를 보면 교정된 z-score값을 볼 수 있습니다.
하지만 decreased/increased (컬러표시)는 activation z-score에 의해서만 조절됩니다.
- 원문 참고 : If the z-score ≤ ‑2, IPA predicts that the process or disease will decrease. Hence the decreased label is displayed. This label is independent of any bias-related corrections. 아래 참고할만한 링크를 첨부하였습니다. For more information on activation z-score and the calculations of bias, please refer the following link: http://pages.ingenuity.com/rs/ingenuity/images/0812%20upstream_regulator_analysis_whitepaper.pdf For more information on the use of downstream effects analysis please refer the following link: http://ingenuity.force.com/ipa/IPATutorials?id=kA250000000TNFbCAO