Chronic kidney disease and pathogen sensors

There are several receptors that detect microbial products. They are responsible for triggering an inflammatory response by the innate immune system. This is important to guard against infections. Sometimes inflammation can happen when it is not needed to protect against infection. A low level of inflammation is sometimes detected in chronic diseases.

Toll-like receptor 4 (TLR4) detects lipopolysaccharides (LPS). The outer layer of gram-negative bacteria contains LPS. TLR4 might also detect some molecules released during tissue damage. We investigated the role of TLR4 in chronic kidney disease (CKD).

Breaking TLR4 so it could not trigger the innate immune system protected against CKD.

Lightning talk slides on deep learning with keras

At the DCPython Office Hours event this month I gave a lightning talk on convolutional neural networks implemented with the keras library. The notebook is now up on github.

Deep neural networks are typically too slow to train on CPUs. Instead, GPUs are used. The example in the notebook uses a relatively small network so should be runnable on any hardware.

Lightning talk slides on web server log analysis with pandas

At the DCPython Office Hours event in May I gave a lightning talk on using pandas to analyse nginx access logs. The notebook is now up on github.

Using sound to guide cells to the kidney

Many diseases are very complex, involving many biological pathways. It is very difficult to find a single drug with actions on enough pathways to slow or reverse many diseases. As an alternative there is excitement around using cells to treat disease. Cells can sense their environment and change their response to match.

Mesenchymal stem cells (MSCs) have performed well in many types of disease. I currently work in the Renal Diagnostics and Therapeutics Unit at NIDDK. Before I joined the group they found MSCs were beneficial in sepsis. Treatment had to be soon after injury limiting the clinical usefulness. One potential reason for this was most MSCs got stuck in the lungs. If more cells got to the kidneys they might work better. At the time it was not known how to guide more cells to the kidneys.

Recently, we have worked with Scott Burks and Joe Frank from the Clinical Center at NIH. They are able to use pulsed focused ultrasound to alter tissues to recruit more MSCs. We tested pulsed focused ultrasound in a cisplatin model. Cisplatin is a treatment for some types of cancer but often causes kidney injury. Pulsed focused ultrasound guided MSC treatment reduced damage from cisplatin.

Comparing biomarkers for acute kidney injury

Acute kidney injury (AKI) is a common complication in hospitalized patients. Patients with AKI are more likely to die or have a worse quality of life after leaving the hospital. Early treatment is our best option for preventing these bad outcomes. To start treatment we first need to detect AKI.

Several biomarkers go up when a patient has AKI. Doctors watch the biomarker levels and start treatment when they rise. There is still uncertainty about which biomarker is best.

I am co-author on a recently published article exploring this issue. We compared creatinine and cystatin C in the serum. Cystatin C predicted kidney function better than creatinine. There was no difference in predicting death.