Adversarial literature synthesis
Your knowledge, your thoughts, your ideas synthesized alongside existing science.
For life science researchers at any career level, from undergraduates through faculty. Outputs are technical by default. Built-in features simplify language when you need them.
The DYNAMIC trial followed patients whose blood showed no detectable tumor DNA after surgery and found their recurrence rates matched patients who received chemotherapy. For a carefully selected subgroup, that's a real argument for stepping back.
I'd push back. The trial wasn't sized to definitively prove equivalence, and it didn't separate patients by the underlying tumor biology. Mismatch-repair-deficient tumors behave differently. Changing how we treat people on this single result feels premature.
Promising direction, not yet decisive. A follow-up trial that explicitly stratifies by tumor biology should come before any clinical change. Patients deserve a more confident answer.
The full debate carries forward: transcript, citations, evidence tiers, unresolved disagreements. An assistant initialized with all of it opens with targeted questions about the critical design gaps.
See it for yourself
The literature doesn't agree with itself.
Your synthesis tool shouldn't either.
Evidence-graded · Citation-weighted · Adversarial by design
The workflow
Every stage shares context with the next. The design emerges from the evidence that was actually debated, not from the researcher's memory of it.
Stage 01
Sharpen your research question with structured guidance before the debate begins.
Stage 02
Two expert perspectives debate your hypothesis across multiple turns. Every claim must be cited. Every verdict must survive scrutiny.
Stage 03
Every cited paper, claim, and verdict from the debate becomes a node in a navigable map. See where the literature converges, where it splits, and which gaps remain.
Stage 04
The consensus verdict carries forward into a structured study design environment. RCT, cohort, translational. Methodology-specific guidance throughout.
Stage 05
NIH R01, R21, foundation, or industry brief. Section-by-section proposal writing grounded in the evidence you actually debated.
Stage 06
Full transcript, design outline, and proposal scaffold exported as structured documents ready for submission or collaboration.
Evidence tagging, consensus checkpoints, and citation-quality weighting via NIH iCite RCR.
The deliverable
When you are ready to write up what you found, the Export builder packages every stage of your work into a customizable report. Scroll through the pages, toggle sections in or out, then ship something a reviewer can actually read.
The composite p16/CD68 marker has biological merit but is currently unsupported by direct human evidence. CD68 alone is a blunt instrument — it counts macrophages without polarization context. Senescent stromal cells expressing p16 may initiate the SASP signal while macrophages amplify it, but the measurement standard for human VAT remains unsolved.
[COHORT] for adipose macrophage burden. [PRECLINICAL] for IL-6 source mechanism. [EXPERT] for polarization continuum. No [RCT] evidence currently available in the indexed corpus.
Whether DXA-derived visceral fat measurement is precise enough in high-BMI populations to serve as the stratification backbone, replacing MRI.
Every PMID cross-checked against the indexed corpus. Off-retrieval citations flagged automatically with a warning marker.
Cohort / Observational
Adults with metabolic syndrome + age-matched controls undergoing elective abdominal surgery
VAT volume by imaging (not BMI), disease duration, exclude statin/metformin/glucocorticoid users
Multiplex immunofluorescence with automated cell-level p16 quantification; CD68 polarization subtyping (M1/M2)
Circulating IL-6 and TNF-α correlated with VAT p16⁺ density and CD68⁺ subtype composition
Aim 1: Test whether VAT p16⁺ cell density predicts circulating IL-6 in metabolic syndrome patients independently of BMI and adipocyte size.
Aim 2: Determine whether CD68⁺ macrophage polarization state moderates the p16/SASP relationship in VAT.
Aim 3: Establish FFPE-vs-fresh-frozen concordance for p16 quantification in adipose tissue across multiple sites.
First human study combining senescence and macrophage burden in metabolic syndrome to predict systemic inflammation, with imaging-based VAT stratification replacing BMI matching.
What makes Colo different
Most research tools answer: what does the literature say?
Colo answers: is this hypothesis worth pursuing, and how?
That requires adversarial reasoning, not summarization. And a workflow
that carries the evidence all the way from debate to proposal.
Reasoning
Evidence quality
Retrieval
Transparency
Continuity
Scope
peer-reviewed abstracts indexed
across 16 biomedical specialties
Two friendly turtles, Lewis and Sage, take turns answering science questions in plain words. Built for the curious, ages 8 to 13.
Try Colo for Kids
→
CD133 promotes multidrug resistance in colorectal cancer via the AKT/NF-κB/MDR1 signaling pathway, representing a potential therapeutic target to overcome chemoresistance.
Meta-analysis of 37 studies confirms CD133 expression correlates with advanced tumor features, worse overall survival, and poorer disease-free survival in colorectal cancer patients.
Acquired resistance mechanisms to anti-EGFR therapy in CRC differ by treatment regimen, with combination chemotherapy markedly reducing emergence of resistance mutations compared to monotherapy.
Pooled analysis of four Italian trials demonstrates anti-EGFR rechallenge provides clinical benefit in refractory ctDNA RAS/BRAF wild-type metastatic colorectal cancer patients.
High ALDH1A1 expression in colorectal cancer associates with poor differentiation, right-sided location, elevated levels in liver metastases, and shorter overall survival.
An international consortium established four consensus molecular subtypes (CMS1–4) providing the most robust, biologically interpretable classification system for colorectal cancer.
An FFPE-optimized CMS classifier improves classification accuracy in clinical CRC samples and confirms CMS2/3 as predictive of anti-EGFR therapy response.
NanoString nCounter is the only platform enabling successful translation of an 18-gene RAS pathway activation signature from fresh-frozen to FFPE colorectal cancer tissues.
Molecular biomarker-based PRESSING selection, not anatomical tumor sidedness, should determine anti-EGFR eligibility in metastatic CRC, enabling biologically accountable treatment decisions.
FOLFOX combined with anti-EGFR monoclonal antibodies significantly improves PFS, OS, and ORR in RAS wild-type left-sided metastatic colorectal cancer patients as first-line treatment.
ProBio is an outcome-adaptive, biomarker-driven randomized platform trial expanded to include both metastatic castration-resistant and de novo metastatic hormone-sensitive prostate cancer patients.
The FAME framework for prospective, adaptive meta-analysis of aggregate trial data reduces bias and optimizes timing for definitive evidence synthesis in systematic reviews.
Cited during the debate but not present in the prebuilt corpus — no embedding similarity available. In a live session, the corpus could be expanded to include it.
Cited during the debate but not present in the prebuilt corpus — no embedding similarity available. In a live session, the corpus could be expanded to include it.
RAS mutations and HER2/MET amplification are the predominant mechanisms of acquired resistance to anti-EGFR therapy in RAS/BRAF wild-type metastatic colorectal cancer.
A multi-society expert panel established 21 evidence-based guideline recommendations for standard molecular biomarker testing to guide EGFR-targeted and chemotherapy decisions in colorectal cancer.