Genome-wide transcription profiling of human sepsis: a systematic review

Table 3 Methodological quality of microarray experiments

	Prucha[14]	Tang-1[15, 16]	Ramilo[17]	Tang-2[18]	Talwar[8]	Payen[19]	Cobb[20, 21]	Pachot[22]	Prabhakar[9]	Calvano[10]	Wong[23–26]	Johnson[27, 28]
Data acquisition
Tissue homogeneity of target samples	Low	High	High	High	High	High	High	Low	High	Low	Low	High
Experiments follow miame criteria^¶	Yes	Yes	Yes	Yes	Not clear	Yes	Not clear	Not clear	Not clear	Not clear	Yes	Not clear
Reporting of normalization method	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes
Analytical issues
Method for gene selection	t test	t test	Non-parametric test	t test	ANOVA	t test	Multiple	Not clear	Not clear	SAM	ANOVA and fold change	Non-parametric test
Issue of variance estimation addressed	No	Yes	No	Yes	No	No	Not clear	Not clear	Not clear	Yes	No	No
Comparison to other diagnostic markers	No	No	No	No	No	NA	Yes	Yes	No	No	No	Yes
Correction for multiple testing	Yes	Yes	Yes	Yes	Yes	NA	Yes	Yes	No	Yes	Yes	Yes
Reporting of classifier performance	No	Yes	No	Yes	NA	NA	No	Yes	NA	NA	Yes	NA
Reporting of prediction accuracy	No	Yes	Yes	Yes	NA	NA	Yes	Yes	NA	NA	Yes	NA
Validation of data
Cross validation of signature genes	No	Yes	Yes	Yes	No	No	Yes	Yes	NA	No	Yes	Yes
External validation in independent samples	No	Yes	Yes	Yes	No	No	Yes	Yes	NA	Yes	Yes	No
Ratio of test/training sample size	NA	1.14	2.00	1.00	NA	NA	0.50	0.23	NA	0.75	0.77	NA
Adjustment for confounders	No	Yes	Yes	No	NA	NA	No	No	No	NA	Yes	Yes
Raw data made publicly available	No	Yes	Yes	Yes	Yes	Yes	No	No	No	No	Yes	No
PCR validation	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	No	Yes

Minimum Information About Microarray Experiment checklist [7]. ANOVA denotes analysis of variance. SAM denotes Significance Analysis of Microarrays [29].
NA denotes not applicable.

ISSN: 1364-8535